International Journal of Advances in Applied Sciences (IJAAS) Volume 7, issue 4, Dec. 2018 by ijaas_journal

ISSN: 2252-8814

IJAAS

International Journal of

Advances in Applied Sciences

Advances in Applied Sciences (IJAAS) is a peer-reviewed and open access journal dedicated to publish significant research findings in the field of applied and theoretical sciences. The journal is designed to serve researchers, developers, professionals, graduate students and others interested in state-of-the art research activities in applied science areas, which cover topics including: chemistry, physics, materials, nanoscience and nanotechnology, mathematics, statistics, geology and earth sciences.

Editor-in-Chief: Qing Wang, National Institute of Advanced Industrial Science and Technology (AIST), Japan Co-Editor-in-Chief: Chen-Yuan Chen, National Pingtung University of Education, Taiwan, Province of China Bensafi Abd-El-Hamid, Abou Bekr Belkaid University of Tlemcen, Algeria Guangming Yao, Clarkson University, United States Habibolla Latifizadeh, Shiraz (SUTECH) University, Iran, Islamic Republic of EL Mahdi Ahmed Haroun, University of Bahri, Sudan

Published by:

Institute of Advanced Engineering and Science (IAES) Website: http://iaescore.com/journals/index.php/IJAAS/ Email: info@iaesjournal.com, IJAAS@iaesjournal.com

Information for Authors International Journal of Advances in Applied Sciences (IJAAS) is an interdisciplinary journal that publishes material on all aspects of applied and theoretical sciences. The journal encompasses a variety of topics, including chemistry, physics, materials, nanoscience and nanotechnology, mathematics, statistics, geology and earth sciences. Submission of a manuscript implies that it contains original work and has not been published or submitted for publication elsewhere. It also implies the transfer of the copyright from the author to the publisher. Authors should include permission to reproduce any previously published material.

Ethics in publishing For information on Ethics in publishing and Ethical guidelines for journal publication (including the necessity to avoid plagiarism and duplicate publication) see http://www.iaescore.com/journals/index.php/IJAAS/about/editorialPolicies#sectionPolicies

Paper Submission You must prepare and submit your papers as word document (DOC or DOCX). For more detailed instructions and IJAAS template please take a look and download at: http://www.iaescore.com/journals/index.php/IJAAS/about/submissions#onlineSubmissions The manuscript will be subjected to a full review procedure and the decision whether to accept it will be taken by the Editor based on the reviews. Manuscript must be submitted through our on-line system: http://www.iaescore.com/journals/index.php/IJAAS/ Once a manuscript has successfully been submitted via the online submission system authors may track the status of their manuscript using the online submission system.

IJAAS

International Journal of

Advances in Applied Sciences

Data Visualization and Analysis of Engineering Educational Statistics Siddhaling Urolagin, Ishita Upadhyaya, Afrah Abbas Thakur

309-316

Mobile Application Development with Android Rajesh Keshavrao Deshmukh, Shilpa Markandey, Pooja Sahu

317-321

Automation of DMPS Manufacturing by Using LabView & PLC F. Fareeza, Chunchu Rambabu, S. Krishnaveni, Abel Chernet Kabiso

322-333

Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus Images Using CBRIR based on Lifting Wavelets S. S. Tadasare, S. S. Pawar

334-346

Recycling of Industrial Waste Water for the Generation of Electricity by Regulating the Flow Control Sensor using IoT Kaviyaraj. R, Karthika M S, Jeni Narayanan L.A, Saleekha

347-352

Coding Schemes for Implementation of Fault Tolerant Parrallel Filter Nutenki Siddhartha, G. Renuka

353-360

Lossless 4D Medical Images Compression Using Adaptive Inter Slices Filtering Leila Belhadef, Zoulikha Mekkakia Maaza

361-368

A Secure Data Transmission Scheme using Asymmetric Semi-Homomorphic Encryption Scheme S. Nagavalli, G. Ramachandran

369-376

Responsibility of the contents rests upon the authors and not upon the publisher or editors.

IJAAS

Vol. 7

No. 4

pp. 309-376

December 2018

ISSN 2252-8814

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 309~316 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp309-316



309

Data Visualization and Analysis of Engineering Educational Statistics Siddhaling Urolagin, Ishita Upadhyaya, Afrah Abbas Thakur Department of Computer Science, Birla Institute of Technology & Science, Pilani, Dubai International Academic City, Dubai, United Arab Emirates

Article Info

ABSTRACT

Article history:

Engineering, is one of the most popular fields of higher education in the modern day world. Majority of the students these days opt for engineering as a career, due to the vastness of choices provided by engineering. Mechanical, Electrical, Computer Science, Civil and Biotechnology are the various disciplines and have varying strength in terms of number of students who join a particular discipline. In this research, we have gather data from various published articles about engineering education and carried out the data visualization and analysis using Tableau 9.2. The objective of the analysis is to help the students to make the decision and the choice about discipline of engineering from which particular university would be the most suitable based on the data collected and represented. Various categories of statistics such as number of graduates from a particular university in a particular discipline, and which university had the maximum number of graduates in a certain year will help the students make their decisions about their future in a more easy and a sorted manner.

Received Mei 12, 2018 Revised Jul 2, 2018 Accepted Aug 12, 2018 Keyword: Data Visualization Engineering Education Statistical Analysis

Corresponding Author: Siddhaling Urolagin, Department of Computer Science, Birla Institute of Technology & Science, Pilani, Dubai International Academic City Dubai, United Arab Emirates. Email: siddhaling@dubai.bits-pilani.ac.in

INTRODUCTION A term which covers several industries and applications broadly is known as “Engineering”. The three fields namely Science, Mathematics and Technology merge together to provide solutions to real-world issues in the form of “Engineering”. Engineering educates students and helps them to develop and making those successful engineers through technical expertise, social awareness and innovation [1]. Among all the professions in the world, Engineering is one of the preferred higher education choice by many students [2]. Due to the increase in number of engineering students, the number of graduates from those universities has been rising annually since 2007. According to a research, a large percentage increase in the enrollment of Computer Science and Electrical Engineering [3] has been seen in the universities all-round the globe [4]. The higher education helps students to have a variety of career options to select and be successful engineers [5]. The Universities are considered to be good on the basis of their quality of education, quality of faculty, research output and per capita performance. In this paper, an effort has been made to analyze the data corresponding to engineering universities such number of candidates taking admission, prospective employments by graduates, number of candidates successfully completing graduation etc. A comparison between different branches of engineering disciple is carried out. Such an analysis is not only useful for students looking to join engineering universities but to the universities also. Students seeking new admission can come to know current trends and opportunities in various engineering discipline. With the help of data analysis universities can define new branches based on trends and improve various aspects of existing branches. The research work is organized as

Journal homepage: http://iaescore.com/online/index.php/IJAAS

310



ISSN: 2252-8814

follows; in section 2 various statistics about engineering education are given, section 3 discusses about number of degrees awarded in various disciples. Engineering branch wise statistical analysis is presented in section 4. In section 5, analysis on female graduated students given and in section 6 salaries of various engineering branches is presented. The section 7 covers conclusion.

ENGINEERING EDUCATION Engineering has some of the major disciplines such as Mechanical, Electrical, Chemical, Biomedical, Computer Science etc., [6] providing solutions to real world problems. Recently several new branches and specializations have evolved showing advancement in modern society. The Computer Science field has been ever evolving since the 1980s [7]. Computer Science has two major fields: Computer Hardware Engineering and Computer Software Engineering. Today, there are several specialties have emerged in field of Computer Science. In the field of Computer Science Engineering, around 4,881 degrees were awarded in U.S. in year 2015. Out of which, 14.8% of the degree holders were female. University of California, San Diego has handed out the highest number of Computer Science degrees in 2015 with 369 graduates. The number of enrollments in that year was 65,244 [8]. Electrical Engineering deals with designing and working with electrical systems, basic requirements to join Electrical Engineering is to have a strong base in Mathematics and Physics. Based on [9] in 2015, 11,385 degrees were handed out across U.S. in this field and 12.5% of which were earned by female graduates. University of Illinois had the highest number of Electrical Engineering graduates in 2015. The enrollment number was 102,519 across U.S. Mechanical Engineering has a vast variety of sub branches and this branch is highly focused on practical applications [9]. In 2015, Mechanical Engineering was the branch with the highest number of graduates in the U.S. with an astounding number of 25,436 [8]. Only 13.2% of these degrees were awarded to women. Maximum number of degrees awarded by Georgia Institute of Technology which of 479 in number. A very high enrollment number is noticed for this branch of engineering in the U.S. alone, with 138,437 students in 2015 [8]. Chemical Engineering mostly deals with the practical applications of chemistry. It involves creating new products with application in pharmaceuticals, cosmetics or the several other fields that involve chemical engineering. In year 2015 around 9,090 Chemical Engineering graduates from the U.S. alone [8]. Ohio State University has awarded the highest number of these degrees and 48,879 students were enrolled into this stream in the U.S. in 2015. As given in [10][11], Computer Hardware Engineers can expect an average annual salary of $110,650 with a median entry level payment of $66,238. Job opportunities in this field are expected to rise by 7.4% by 2022 [10][11]. Software system developers, on the other hand can expect an average annual salary of $106,050. The median entry level salary for the same is $68,510. Jobs in this field may rise up by 20.4% [12]. An Electrical Engineer receives an average annual salary of $97,340 with a median entry level payment of $64,981. Mechanical Engineering graduates can expect a mean annual salary of $88,190. The median entry level payment for this field is $62,527. Jobs for Mechanical Engineers are expected to rise by 4.5% by the year 2022 [11]. In the field of Chemical Engineering, one can expect an average annual salary of $103,960. The median entry level payment is $67,006 [10].

ENGINEERING DEGREES AWARDED The Figure 1. depicts the number of bachelor degrees awarded by various disciplines from different U.S universities, with Mechanical discipline awarding highest number of bachelor degrees of 25,436 followed by Civil discipline with 11,900. Computer Science discipline at the fourth place with 10,970 degrees being awarded. In Mining discipline 335 number of bachelor’s degree being given.

IJAAS Vol. 7, No. 4, December 2018: 309 – 316

IJAAS

ISSN: 2252-8814



311

Figure 1. Bachelor’s degrees awarded by various disciplines.

The number of degrees awarded in comparison with total number of enrollment across various U.S universities is depicted in Figure 2. Arizona State University has highest number of enrollment of 11,572 and second highest is Texas A&M University with 10,541 as shown in Figure 2. The number of degrees awarded by Arizona State University is 1,123 and for Texas A&M University is 1,490. Highest number of degree awarded by University of Illinois with 2,012.

Figure 2. Number of degrees awarded and enrollment number of universities. Data Visualization and Analysis of Engineering Educational Statistics (Siddhaling Urolagin)

312



ISSN: 2252-8814

Figure 3(a) Enrollment numbers and degrees awarded by disciplines. (b) Number of enrollment and degree awarded. In Figure 3(a). Number of students enrolled and number of degree awarded for different engineering disciplines are shown. Mechanical discipline has the highest number of enrollments as 138,437 and the highest number of bachelor degrees being awarded as 25,436. Aerospace discipline has nearly as much enrollments as Industrial/Manufacturing/Systems discipline, a total of 20,331. Architectural discipline had the lowest enrollments and bachelor degrees being awarded as 3,135 and 568 respectively. The comparison between enrollment and degree awarded in Computer Engineering for Drexel University, Iowa University, Purdue University, Virginia Tech and West Virginia University are show in Figure 3(b). Iowa and Purdue Universities have highest number of enrollment as 8,209 while West Virginia University has lowest as 3,965. While number of bachelor degrees awarded by Purdue is 1,626 and West Virginia as 36 in Computer Engineering.

ENGINEERING BRANCHES The number of degree awarded in Mechanical Engineering is depicted using tree map in Figure.4. Georgia Institute of Technology has granted the most number of degrees in the Mechanical Engineering as 479 followed by it Purdue University as 413. The California State University has awarded lowest number of degrees as 154. The bubbles representation is depicted in Figure. 5(a) and (b). In Figure. 5(a) shows number of degrees awarded in Computer Engineering by various universities. University of California, Berkeley has awarded maximum number of degrees as 283. Number of degrees awarded in Computer Science (outside engineering) is shown in Figure. 5(b) University of California, Irvine followed by University of Texas, Austin have awarded as 343 and 376 respectively.

IJAAS Vol. 7, No. 4, December 2018: 309 – 316

IJAAS

ISSN: 2252-8814



313

Figure 4. Number of degrees awarded in Mechanical by different universities.

Figure. 5 (a) Number of degrees awarded by universities in Computer Science. (b) Number of degrees awarded by in Computer Science (outside engineering).

Data Visualization and Analysis of Engineering Educational Statistics (Siddhaling Urolagin)

314



ISSN: 2252-8814

Figure. 6(a) Number of degrees awarded in Aerospace Engineering. (b) Number of degrees awarded in Civil Engineering by various universities. Highest seven universities based on number of degrees awarded in Aerospace and Civil Engineering are depicted in Figure 6(a) and Figure 6(b). Embry Riddle Aerospace University tops the list with a total of 202 degrees being awarded. The California State Polytechnic University awarding a total of 209 degrees and Texas A&M University seconds the list by awarding a total number of 205 degrees in Civil Engineering.

FEMALE ENGINEERING STUDENTS In Figure. 7 we have analyzed percentage of degrees awarded to women in various engineering disciplines. Biological and Agricultural Engineering percentage of 49.7, which is the highest percentage of degrees awarded to women followed by Biomedical Engineering of percentage 31.2%. The percentage of degree earned by women in Computer Science is 14.8%, in the Mechanical Engineering it is 13.2%.

Figure 7. Percentage of degrees awarded in various disciplines for women. IJAAS Vol. 7, No. 4, December 2018: 309 – 316

IJAAS ISSN: 2252-8814  315 6. SALARIES In Figure. 8 a comparison of annual salary for various disciplines is presented. Three main matrices are shown: median entry level salary, mean annual salary and top 10% of salary in various engineering discipline. Geological and Mining engineering is highest in terms of top10% salary structure. Based on mean annual salary, Computer Hardware Engineering noted to be highest with $114970. Mechanical Engineering has mean annual salary of $88190.

Figure 8. Annual salary for various disciplines.

CONCLUSION Based on analysis carried out in this research with the help of Tableau 9.2, Mechanical Engineering has the highest number of degrees awarded and enrollment numbers. Certain engineering branches such as Mining, Environmental Engineering, Engineering Management, Engineering Science, Architectural Engineering and Nuclear Engineering are available in few colleges and have very few students have purse their career. Certain branches of universities are far more successful than their other branches. Some universities have very high enrollment numbers but are lower in rank when it comes to the total number of graduates. For instance Mechanical Engineering has the highest enrollment numbers as well as the highest number of degrees awarded. A large percentage of women opt for Biology related fields such Biological and Agricultural Engineering and Biomedical Engineering. University of Illinois, Urbana-Champaign has the highest total number of degrees awarded and is one of the top most choices of students interested in pursuing engineering. Computer Engineering has been observed to have very high enrollment numbers as well as the highest median entry-level salary and mean annual salary. Based on statistical analysis Computer Science can be seen as very promising field of engineering. Despite having the highest enrollment numbers and number of degrees awarded, Mechanical Engineering is not very rewarding in terms of salary. In terms of salaries, the Computer Engineering field is very promising for the graduates. Chemical Engineering pays high entry-level salaries as well. Despite very poor enrollment numbers into Geological and Mining Engineering, the top 10 percent of people working in this sector earn the highest salaries when compared to their counterparts in fields such Material Science Engineering, Environmental Engineering, Civil Engineering, Biomedical Engineering. This could probably be because of a high demand and low supply of such engineers. Environmental and Civil Engineers have very low median entry-level salaries compared to other branches

Data Visualization and Analysis of Engineering Educational Statistics (Siddhaling Urolagin)

316



ISSN: 2252-8814

REFERENCES Edward F. Crawley, Johan Malmqvist, Sören Östlund, Doris R. Brodeur, Kristina Edström. “Rethinking engineering education”, The CDIO Approach, New York, Springer, 2014. [2] John Heywood, "Engineering Education", IEEE press Wiley Interscience, 2005. [3] Mark Guzdial, “Generation CS' Drives Growth in Enrollments”, Communications of the ACM, Vol. 60 No. 7, Pages 10-11, 2017. [4] Heike Jöns, Michael Hoyler, “Global geographies of higher education: The perspective of world university rankings”, Geoforum, vol.46, pp 45-59, 2013. [5] National Science Foundation, Science and Engineering Indicators 2012: Overview, National Science Foundation, Arlington, VA, 2012. [6] Froyd, J.E.; Wankat, P.C.; Smith, K.A., “Five Major Shifts in 100 Years of Engineering Education”, Proceedings of the IEEE. Vol. 100, issue 13, pp. 1344-1360, 2012. [7] Ceruzzi, Paul E. “A History of a Modern Computing” The MIT Press. 2003. [8] Yoder, Brian L. “Engineering by the Numbers”, In American Society for Engineering Education. 2012. [9] Dym, C.L.; Agogino, A.M; Eris, O.; Frey, D.D.; Leifer, L.J. , “Engineering Design Thinking, Teaching, and Learning”, in Journal of Engineering Education, vol. 94, issue 1, pp. 103–120, 2012. [10] Engineering Salary Statistics Retrieved from Michigan Technological University, http://www.mtu.edu/engineering/outreach/welcome/salary/, accessed on 5 May, 2017. [11] The Engineering Income and Salary Survey Standard Report, ASME, 2012. [12] Dym, C.L., Agogino, A.M, Eris, O., Frey, D.D., Leifer, L.J, “Engineering Design Thinking, Teaching, and Learning”, in Journal of Engineering Education, vol. 94, issue 1, pp. 103–120, 2012. [1]

IJAAS Vol. 7, No. 4, December 2018: 309 – 316

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 317~321 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp317-321



317

Mobile Application Development with Android Rajesh Keshavrao Deshmukh, Shilpa Markandey, Pooja Sahu Department of CSE, SSIPMT, Old Dhamtari Road, Sejabahar P.O., Mujgahan, Raipur, Chhattisgarh 492015, India

Article Info

ABSTRACT

Article history:

The Android is mobile platform. It is an open source and free operating system application, by Google it is developed and maintained. It was designed essentially for touch screen mobile devices, such as and tablet, computers, smart phones, watch television, cars etc. Android is one of the most widely used mobile OS. Android is a not only operating system but also key applications and middleware. Android is an open source operating system. It is developed by the open handset Alliance, led by Google, and other companies. Those are used to android studio 2.2.3 version and development the mobile application.

Received Feb 9, 2018 Revised Jul 9, 2018 Accepted Aug 11, 2018 Keyword: Android Architecture Security Version

Corresponding Author: Rajesh Keshavrao Deshmukh, Department of CSE, SSIPMT, Old Dhamtari Road, Sejabahar P.O., Mujgahan, Raipur, Chhattisgarh 492015, India. Email: r.deshmukh@ssipmt.com

INTRODUCTION The Android is technology. It is open source operating system, which means that anyone who wants to use Android can do so by downloading the full Android source code. That was used for development mobile based application and desktop based application. This is complete set of software for mobile device such us tablet computer ,smart phone, notebooks ,watch and book readers, set-top boxes etc. it content three things Operating system, middleware, key apps .The android is Linux based operating system. More than 30 companies working to open up the mobile handset environment. Mobile application development is very good demand in market. The application are usually development the java language using the android development tools .the development tools are available in market such as native development kit for application or extension in C or C++, Google app inventor etc. It can run on many different devices from many different manufacturers. The android is powerful and very popular operating system it can supports large number of applications in touch screen Phone. The hardware supports android software which is based on ARM architecture platform. These applications are more comfort Zone and advance Level for the users. The android is an open source operating system that means it’s free and users can easily access and use it the android application. The android has got millions of apps available. The android help us to manage our life one or other way and the application are available in market low cost at that reasons android is very popular in world.

HISTORY The world most popular mobile operating system launch in 2003 Initially, android incorporation founded by Andy Rubin in Palo Alto, California, United States in October, 2003. In 17th August 2005, Google purchase android Incorporation. Since then, android become subsidiary of Google Incorporation. Rich Miner, Andy Rubin, Nick Sears and Chris White is the key employees of Android Incorporation. At that time camera

Journal homepage: http://iaescore.com/online/index.php/IJAAS

318



ISSN: 2252-8814

is more expensive to buy so that android shifted camera into smart phones. Andy Rubin love Robots so that he given the nick name of Android is coworkers. In 2007, Google announces the development of android OS. In 2008, HTC launched the first android mobile. 2.1. Android versions: There are Many Versions of android but Versions 1.0 and 1.1 were not released under specific codename

Table 1. Versions of Android Version

Code name

API level

Release date

1.0

No codename

1.1 1.5 1.6

No codename Cupcake Donut

2 3 4

2.1 2.2

Éclair Froyo

5-7 8

September 23, 2008 February 9, 2009 April 27, 2009 September 15, 2009 October 26, 2009 May 20,2010

2.3

Gingerbread

9-10

December 6,2010

3.1-3.3

Honeycomb

11-13

February 22,2011

4.0

Ice Cream Sandwich

14-15

October 18,2011

4.1-4.2- 4.3

Jelly Bean

16-18

4.4

KitKat

19-20

July 9, 2012 October 312013

5.0

Lollipop

21-22

November 12,2014

6.0

Marshmallow

October 5,2015

7.0

Nougat

24-24

August 22,2016

8.0-8.1

Oreo

26-27

August 21,2017

Figure 1. Of Android Versions

3.     

FEATURE OF ANDROID Storage: A lightweight relational database SQLite, for purposes of data storage it is used. Connectivity: It support many connectivity technology like WIFI, BLUETOOTH, UMTS, WIXMAX, CDMA, and GSM/EDGH. Messaging: The android is support feature of the both MMS and SMS. Web browser: Coupled with Chrome's V8 JavaScript engine supporting CSS3MMS and HTML5, based on the open-source Web Kit layout engine. Multi-touch: Android have a feature of Multitouch which is first made by handset HTC Hero.

IJAAS Vol. 7, No. 3, September 2018: xx – xx

IJAAS ISSN: 2252-8814  319  Multi-tasking: Various application can run simultaneously. The user can jump from one task to another task at the same time.  Tethering: The android is Supports sharing of Internet connections as a wired/wireless hotspot.

4.     

ANDROID ARCHITECTURE The software stack is spilt into five layer: The application layer The application frame works The libraries The runtime The kernel

Figure 2. Android Architecture

4.1. The Kernel The android is based on Linux 2.6 kernel architecture. This layer is core of android architecture. It provides service like processor and power management, security, memory management etc. 4.2. The Native Libraries There Android has its own libraries, which is written in C/C++. These libraries cannot be accessed directly. Using application framework we can access these libraries. There are libraries which help to access many type of web browsers for this there is a web library and libraries for video and android formats etc. 4.3. The Android Run Time The Android Runtime is an application. Which is run time environment .It is used by operating of android. Where you have limited battery, limited CPU, limited memory. Android have its own virtual machine is called Dalvik. It is run on android application. Dalvik is used in mobile devices such as tablets, phone etc. Program are commonly written in java and compile to byte code. 4.3.1. Android Run Time This is in blue meaning that it’s written in the java programming language. The internal library contains all of the collection utilities, class, IO and all the utilities. 4.4. Appication Framework The application framework is the toolkit that all application use and this is all written in java programming language These applications include the ones that come with a phone like the phone application, Mobile Application Development with Android (Rajesh Keshavrao Deshmukh)

320



ISSN: 2252-8814

or home applications, .It includes application written by Google, and includes apps that will be written by you. Every application use same APIs and framework. These are as follows:  Activity manger: the manages the lifecycle of application. It enables proper management of all activities. All the activities are controlled by manager.  Resource manager: it provide access to non-code resources such as graphic etc.  Location manager: if fires alerts when user enter or leaves a specified geographical location  Package manager: it is use to retrieve the data about installed packages on device.  Window manager: it is use to create view and layouts.  Telephony manager: it is use to handle setting of network connection and all information about services on device. 4.5. Application Layer Application layer is in the top layer. In this layer sharing of data and transfer of files. It involves the contacts application, the home application, apps and the browser. It is the top layer in android architecture. All the native application are like Google maps, camera, SMS, browser, calendars, and contacts are. With the help of application of framework to operate this application work with end user

SECURITY All Android is design more technical security features and simply presented to user that means they can be easily controlled through the interface. To improving your Android device’s security this method is straightforward methods can include: using a pin or fingerprint, password, setting your phone to lock after a period of inactivity, only enabling wireless connections that you use, and only installing Android apps you believe and have personally cheked. Google allows Android applications into its marketplace and this application are tested and proven, secure by Google, it means that the user will be protected and less of a chance of installing virus and malicious app. For installation of an application, Android security system prompts the user to allow this installation. Meaning that it is impossible to remotely install and run an application. By regularly installing system updates users can secure their android devices from attackers. Android is a multi-process system in which each application (and part of the system) runs in its own process. The security between the system and application is getting through standard Linux facilities, such as group IDs and user that are assigned to the applications. When attackers attack on device, android platform help to protect the smart from attackers. There are key components of android security which are described as follows:  Design review: by the developers design of security model is reviewed. So that while using this model the risk level will be very less.  Code review and penetrating testing: the main aim of code review is that it will be test that how the system will become powerful?  Open source and community review: android have open source and community review such as Linux kernel.

ADVANTAGE AND DISADVANTAGE OF ANDROIDS Many advantage or disadvantage of android application.

6.1.         

Advantages of Android Android phone is easy to access. It is the easy to carry. It is more secure. With the support of many application, the user can change the screen display. It is platform independent. The major advantage is android is multitasking. It is support the Google service. It given you better notification. A more mature platform.

IJAAS Vol. 7, No. 3, September 2018: xx – xx

IJAAS ISSN: 2252-8814 6.2. Disadvantage of Android  Some time attackers can attack android.  Android operating system having lot of process because of this android OS hang.  If memory is full then android phone getting slow.  Extremely inconsistence in design among apps.  It is unstable and some time crash.



321

CONCLUSION AND FUTURE SCOPE Android is open source platform it can easily development the mobile based application All the APIs are available in order to develop these applications. It is very secure and its help to protect from attackers. Many version of android s available and latest version is Oreo 8.1. In the field of artificial intelligent many type of beneficial robot can be made which is require in future. Which is in progress to use in cars, watches and also Android TV.

REFERENCES [1]

[2] [3] [4] [5] [6]

Vikas Agrawal, Shruti Agrawal, Rajesh Deshmukh “Analysis and Review of Encryption and Decryption for Secure Communication” International Journal of Scientific Engineering and Research (IJSER) ISSN (Online): 2347-3878, Volume 2 Issue 2, February 2014. Li Ma, Lei Gu and Jin Wang “Research and Development of Mobile Application for Android Platform” International journal of Multimedia and Ubiquitous Engineering. https://www.javapoint.com/androoid-history–and-versions https://www.digitaltrends.com/mobile/android-version-history https://www.scribd.com/document/79741205/Paper-Presentation-on-Andriod http://www.android.com

Mobile Application Development with Android (Rajesh Keshavrao Deshmukh)

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 322~333 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp322-333



322

Automation of DMPS Manufacturing by using LabView & PLC F. Fareeza1, Chunchu Rambabu2, S. Krishnaveni3, Abel Chernet Kabiso4 1,3Dr.

M.G.R. Educational & Research Institute, Chennai, India Minch University, Arba Minch 21, Ethiopia

2,4Arba

Article Info

ABSTRACT

Article history:

This Paper is to enable the Siemens (Programmable Logic Control) CPU 3135A to communicate with the Lab VIEW and to control the process accuracy by image processing. The communication between CPU 313-5A and Lab VIEW is via OPC (OLE for Process Control).Process Accuracy is achieved with the use of Labview Image Processing and Gray Scale matching Pattern. Accuracy in the gray scale matching will purely depend on the calibration of the camera with respect to the corresponding image. The digital output from the labview is communicated to PLC via Ethernet Protocol for the industrial process control. With the use of Labview the dead time while using the normal image vision module in PLC can be minimized. Labview uses the gray scale matching technique which is more accurate than the normal image vision module used in PLC

Received Apr 24, 2018 Revised Jun, 2018 Accepted Jul 28, 2018 Keyword: Camera Ethernet LabView PLC Process Control

Corresponding Author: F. Fareeza, Dr. M.G.R. Educational & Research Institute, Chennai, India. Email: fareezafayaz@gmail.com

INTRODUCTION To automate the DMPS mixing station for achieving better accuracy in the process with the use of Labview Image Acquisition module and PLC Machine vision is a broad term. Many mathematical theories, image acquisition, image processing and analysis, etc. form the whole machine vision world. Therefore, to define machine vision is a difficult task, if these things are considered. Pages and pages could be written on these different topics. For this reason, the main concentration of this thesis document was to develop a simple machine vision application using software readily available in the market. The software used in this thesis paper was Lab- VIEW. However, the basic requirements to understand a machine vision system are discussed in the forthcoming sections.

LITERATURE SURVEY It is reported that mere physical examination and Laboratory shambling report is not helpful in the Process Accuracy, The reason behind it was the atmospheric conditions especially the temperature and the moister content may affect the Process accuracy Anjali Setal (2014), discussed about the various advantages on Labview-PLC based process control for a small statin with digital inputs via Modbus communication Protocol [1]. Forsyth et al (2009) contributed a method where measure of Image Color in gray scale pattern which is more accurate than Color matching Technique by Comparison with a reference image [2]. Papadopoulos, E et al (2008), described a novel approach to automatean industrial process by the Image matching Technique with the use of separate image acquisition module [3]. Siemens Industrial Automation, (2008) explained the operation of the CPU 315- A Series type of Programmable controllers [4]. N. N. Barsoum, et al (2011) explains remote control applications over a wide area which is commonly used in industries these days. Ethernet module was used for achieving remote control [5]. Nargalkar Akshay, et al (2005), is to enable the Allen Bradly PLC (Programmable Logic Control) SLC-500 to communicate with the Journal homepage: http://iaescore.com/online/index.php/IJAAS

IJAAS ISSN: 2252-8814  323 Lab VIEW. The communication between SLC-500 and Lab VIEW is via OPC (OLE for Process Control). Development of OPC using RS Linux OPC Server. OPC is an industry standard provides real time plug-andplay software technology for process control and factory automation [6]. Fairchild, M. (2005). Proposed a new feature of Color matching Technique by Comparison with a reference image [7].

DMPS DMPS (Dimethoxic Proponic acid) is mixed gradually under controlled temperature with Edible Oil (Sunflower) for better quality. For this process 12 Ton of Edible Oil-Sunflower is taken for a batch, in its 12Kg of Sunflower oil concentrated with 120 Grams of DMPS solution is mixed gradually under controlled temperature. For the preparation of DMPS concentrated Sunflower oil 12kg of oil is taken in a vessel (small tank) and it was maintained at 650 C with the use of heater coils then DMPS Solution is dosed gradually to it with the use of dosing pump. During this process the colour of the sunflower oil in the tank gets changed due the concentration of DMPS. When the Colour of the DMPS Concentrated Sunflower Oil gets changed to Lime colour the camera will sense it and sends the signals to Lab view .The command from the lab view will be communicated to PLC-Simens CPU 315-A which governs the whole process with the use of ladder logic programmed in it. For the preparation of DMPS concentrated Sunflower oil 12kg of oil is taken in a vessel (small tank) and it was maintained at 650 C with the use of heater coils then DMPS Solution is dosed gradually to it with the use of dosing pump the control of the dosing pump, Agitator and dosing pump is governed by the PLC Omron-Cpm2a.Temperature control of the vessel is controlled locally with the use of temperature controller with signals from temperature transmitter connected with RTD-PT100.The status signal of the temperature controller is given to the PLC Omron-Cpm2a for the control of the process. As the concentration of DMPS increases the color of the sunflower Oil changes to lime color which is sensed with the use of camera and the image is processed in labview and the status signal is shared to PLC for further action via OPC server of labview which is received by PLC via OPC client.

Figure 1. Block Diagram

Automation of DMPSManufacturing by using LabView & PLC (F. Fareeza)

324



ISSN: 2252-8814

Figure 2. Process Diagram

3.1. Process Sequence 1. Push button ON-):Agitator, Dosing Pump and Heater gets poweredheater uses separate temperature controller loop uses temperature Controller 2. Camera Starts sensing and when it reaches the required colour it passes the signal to plc via Lab view 3. When Image and temperature satisfies condition agitator and dosing pump stops and the outlet actuator starts dropping 3.2. 1. 2. 3.

Techniques Used PLC –Siemens CPU 315-A uses Ladder Logic Programming with Step-7 Logic LabView Uses Graphical programming PLC-LabView Communication is with the use of LabView OPC Server and Omron CX-One Server

COMPONENTS OF MACHINE VISIONSYSTEM A machine vision system typically consists of machine visionsoftware (machine vision tools) and a camera (image acquisition device). Butmany otherthingsneedtobeconsideredforamachinevisionsystem.Eachof the components has its own significance. So none of the componentscan be isolated or segregated. The important components are discussed inbrief in the followingsub-sections. 4.1. Illumination Illumination refers to the light sources that are available around the object being analyzed. It is significant that the object(s) under analysis be clearly visible to the image acquisition device. It ensures that much of the information is retained in the acquired image, and no much image processing needs to be done; thus making the machine vision application simpler to develop. Illuminating object(s) does not mean availability of huge amount of light around the object; it refers the lights to be adjusted in a proper way. Proper illuminationinvolves the right intensity and correct direction of light. It should be done in a way that shadow formation is checked and maximum contrasts can be achieved from the region of interest of the object(s). (Mo- vimed custom imaging solutions, 2007.) The light sources may be fluorescent lights or LED lights or halogen lights, etc. LED lights are more preferred over the other types of light sources, because of their long life and less energy consumption. Depending upon the arrangement of lights, illumination can be direct or indirect.

IJAAS Vol. 7, No. 4, December 2018: 322 – 333

IJAAS ISSN: 2252-8814  325 4.2. Imageacquisition Image acquisition is the most important part in a machine vision system. It involves capturing an image of the object to be analyzed with the help of camera. Different types of cameras can be used for image acquisition; they can include an ordinary mobile camera, a typical digital camera, or even a webcam. But cameras that are tailored specially for industrial use are also available. Depending on the sensor technology used, different cameras can be classified into two categories as follows:  CMOS cameras  CCD cameras “The sensors could be matrix sensors or line sensors. An image sensor converts an optical image into an electronic signal” Choosing a machine vision camera can be a difficult task. However, resolution, sensitivity, and type of cameramonochrome or color, should be considered when buying one. (ALLIED Vision Technologies GMBH, 2006.) Also, the interface the camera uses for communication should be considered. The available interfaces include USB, Ethernet, Firewire, etc. 4.3 CCD Cameras A CCD (Charge-coupled Device) camera uses the CCD sensor technology. The main features of these sensors (and hence cameras) are listed as follows. (Vision Systems Oy. 2012.) The most common camera sensors In-coming charges are stored Equivalent to films of traditional film cameras. Consist of pixels with a typical size of 10µm x 10µm.  These are both color and monochrome.  These are light-sensitive diode sensors.  Each pixel has a micro-lens for focus- ing the light into the sensor surface.  Disadvantage of these sensors is the possibility of over exposure.

Figure 3. CCD Camera

4.4. Depth of Field Depth of Field (DOF) is the measure of the distance between the nearest and the farthest objects in a scene that can be captured by the camera and be acceptably sharp in the image 4.5. Construction of the Station The system consists of cylinders, sensors, a motor as the physical components. As the PLC used in the station requires additional signal modules for the I/O (inputs and outputs) devices, signal modules from Beckhoff (www.beckhoff.com) was used. 4.6. System Communication The PLC was interfaced to the computer using S7 MPI adapter. The bus coupler, LC3100 was networked with the PLC using PROFIBUS cable. Also OPC communication was done using the same S7 MPI adapter. A di- agrammatic representation of communication among the system components is shown in Figure 4.

Automation of DMPSManufacturing by using LabView & PLC (F. Fareeza)

326



ISSN: 2252-8814

LabVIEW

STEP 7

MPI

CPU315-2DP

OPC

Profibus

LC3100

Figure 4. Communications among STEP 7, LabVIEW and CPU315-2DP and LC3100

4.7. STEP 7 Professional STEP 7 PROFESSIONAL (STEP 7 in short) is automation softwarefrom Siemens Industry (www.siemens.com). It is used for program mingsimatic PLC stations. Figure 28 shows the system manger window of simatic STEP 7, where a created paper with some blocks is also shown

Figure 5. Screenshot of Simatic Manager (STEP 7) STEP7 provides provision to code the program using seven differentlan-guages. Some of them are FBD (Function Block Diagram), LD (Ladder Diagram), and SCL (StatementList), etc. The user can freely choose the language. An STEP 7 auto mation program may contain functions (FCs), function blocks (FBs), organization blocks (OBs), sequence functional charts (SFCs), etc. But, every STEP 7 program must have OB1, becauseit is the main function. Details about the STEP 7 programs are not coveredin this thesis. The language used for the automation of the system described in this thesis was FBD

IJAAS Vol. 7, No. 4, December 2018: 322 – 333

IJAAS

ISSN: 2252-8814



327

Figure 6. FBD codes in STEP 7

4.8. CPU315-2DP CPU315-2DP belongs to Simatic 300 group of PLC controllers. It consists of different indicators, a program manipulation key, and a memory card slot. It has 2 serial-connection ports; one of them is for interfacing between STEP7 and the PLC whereas the other is for connecting to a distributed module. Figure 30 alongside shows CPU315- 2DP

Figure 7. CPU315-2DP

4.9. Hardware Configuration Hardware configuration (HW) needs to be done before any program can be downloaded into the PLC. In the hardware configuration, the type of power supply, the CPU model (such as CPU315-2DP), the signal modules and (or) other distributed modules being used need to be specified sothat the software and the hardware can inter-connect. Figure 8 shows the hardware configuration done for this paper

Automation of DMPSManufacturing by using LabView & PLC (F. Fareeza)

328



ISSN: 2252-8814

Figure 8. An example Hardware configuration for simatic PLC

RESULTS AND DISCUSSION

Figure 9. Labview Font Panel

IJAAS Vol. 7, No. 4, December 2018: 322 – 333

IJAAS

ISSN: 2252-8814



329

Figure 10. Step 7 Ladder and Simulation Module

The output terminal from the Unbundleby Name VI was once again wired to Array-to-Cluster VI, whose out put terminal was also once again wired to another Unbundleby Name VI. This time it was en larged to get 4 output terminals. The first output terminal was assigned to ‘Bounding Box [0]>x’; the second to ‘Bounding Box [1]>y’; thethird to‘BoundingBox [2]>x’; and the fourth to‘BoundingBox [3]>y’as shown in Figure11

Figure 12. Changing the cluster size Automation of DMPSManufacturing by using LabView & PLC (F. Fareeza)

330



ISSN: 2252-8814

Figure 13. PLC Ladder & Simulation Module The output terminal from that final Array-to-Cluster VI was connected to the ‘Rectangle’ input terminal of the ‘IMAQ Overlay Rectangle’VI. Also, the input terminal ‘Image’of this VI was wired to the‘Image out’ output terminal of the vision assistant. Finally, the output terminal from the Overlay Rectangle VI was wired to the input terminal of the ‘Image Display’VI. 5.1. Method of Interfacing CONFIGURING PLC DRIVER Here we choose RS-232 DF1 devices and click on Add New. Next we name the PLC and then start the configuration in Configure RS-232 DF1 devices. CREATING NEW OPC TOPIC Create a new OPC Topic and save it for further use. All the inputs and outputs that are available in the PLC will be available under topic which we have created

CONCLUSION From the paper performed, it was obvious that machine vision technology and NI Vision tools together can be used for sorting objects in a factory production line. It was confirmed that out of many machine vision tools, the pattern-matching algorithm could be applied for the object sorting purpose. Also, it was clear that a simple webcam could be used for performing machine visions tasks. Because, a webcam was used for pattern matching, it is not for sure that it can be used for different machine vision applications. But, if the application is independent of different image processing tasks, it might be possible to use webcams instead of other industrial cameras. The pattern matching was done at different times of the day; every time the score generated by the vision assistant needed adjustment for the correct pattern matching. This shows that pattern matching is based on direct correlation matrix algorithm. The score generated for the image template was 900. If the score was increased over 900, none of the objects matched the pattern; below 900 pattern matched, but if the score was below certain score (for example 850), the objects which tend to be of the same pattern also matched. In that case, the aim of pattern matching failed. Although pattern matching is not affected by lighting, it was seen that shadow formation of some other objects over the analyzed object hindered to match the pattern stored in the pattern template. So to say that pattern matching is independent of lighting might be somewhat unrealistic. It might be said that pattern

IJAAS Vol. 7, No. 4, December 2018: 322 – 333

IJAAS ISSN: 2252-8814  331 matching not affected by lighting directly, but there might be some other indirect cause relating to light that affects pattern matching. It was also confirmed that PLC could be integrated with LabVIEW using OPC communication protocol; communication was enabled using NI OPC Servers in this thesis. The integration of simatic PLC was illustrated in this thesis but almost all the PLCs can be integrated with LabVIEW using NI OPC servers; provided there is an OPC driver for the PLC. In this thesis document it was possible to present only a simple pattern recognition technique for the object sorting. Although the objective of sorting object using machine vision tool was met, it was realized that geometric matching tool would be more robust for sorting objects. The objects used in the paper were all similar; defining only one parameter could be enough in such condition to get the result. But the case would not be the same always. The objects may have different shapes and sizes; in such cases, it is not enough to match only the pattern of the object. Measuring distances between two points, measuring diameters, etc. could be added to the application to make it more powerful and error-free. The main idea was to interface LabVIEW and PLC for more effective and efficient process control. This paper of interfacing two most powerful technologies ruling the industries lead to many new features like acquiring data at faster rate, controlling the process accurately, providing Multi-Tasking operations, Remote controlling and several other features. Used correctly, multithreading offers numerous benefits including more efficient CPU use, better system reliability, and improved performance on multiprocessor computers. Using LabVIEW, we can start today to maximize performance on multithreaded operating systems and multiprocessor computers without increasing either your development time or the complexity of your application. Because the multithreading technology of LabVIEW is implemented transparently, no extra programming is required to take full advantage of multithreading technologies.

ACKNOWLEDGEMENTS We, authors of this paper would like to show our gratitude to our family, friends and colleagues for sharing their pearls of wisdom with us during the course of this research paper. REFERENCES [1]

[2] [3] [4] [5] [6]

[7] [8]

[9] [10] [11] [12] [13] [14] [15] [16] [17]

Anjali S,et al(2014)., Application of MODBUS to Communicate the PLC and Lab VIEW for Real Time Process Control, International Journal of Emerging Science and Engineering (IJESE) ISSN: 2319 – 6378, Volume-1, Issue11, September 2013 Forsyth, et al (2009) Computer Vision: A Modern Approach, Papadopoulos, E et al (2008) Pattern Analysis and Machine Intelligence Siemens Industrial Automation, (2008) “CPU 315-A Series Programmable Controller,” Simens, Technique Report W393-E1-14. N. N. Barsoum, et al (2011) “Ethernet Control AC Motor via PLC Using LabVIEW, Intelligent Control and Automation, 2011, 2, 330-339. Nargalkar Akshay, et al (2005) “Real Time Automated Control of Industrial Processes with PLC –LABVIEW Communication”, International Journal for Research in Science & Advanced Technologies ISSN: 2319-2690 Issue1, Volume-1, pp 035Fairchild, M. (2005). Color ApperanceModeles. Chichester, UK: Wiley-IS&T proposed a new feature of Color matching Technique by Comparison with a reference image. ALLIED Vision Technologies 2006. Application Note: Choosing a Machine Vision Camera. Accessed 15th March 2013.http://www.alliedvisiontec.com/fileadmin/content/PDF/Support/Application_Notes/ALL_ApplNote_Choosin g_A_Machine_Vision_Camera.pdf Basic Photography Tutorials 2001. An Introduction to Exposure. Accessed 16th March 2013.http://www.silverlight.co.uk/tutorials/compose_expose/exposure.html Davies, E. 2012. Computer and Machine Vision. 4th edition. Waltham: Academic Press. EURESYS 2005. ADR Technology. Accessed 18th March 2013. http://www.euresys.com/Products/grablink/ADRTechnology.asp Gonzalez, R., Woods, R. &Eddins, S. 2009. Digital Image Processing using Matlab. 2nd edition. Gatesmark Publishing. Logitech 2010. Logitech HD WebCam C615. Accessed 15th March 2013. http://www.logitech.com/enus/product/hd-webcam-c615 Lode’s Computer Graphics Tutorial 2004. Image Filtering. Accessed 30th April 2013. http://lodev.org/cgtutor/filtering.html Machine Vision, pdf-file (ebook). Accessed 15th January-23rd April 2013. http://www.cse.usf.edu/~r1k/MachineVisionBook/MachineVision.pdf Movimed Custom Imaging Solutions 2007. Illumination. Accessed 16th April 2013. http://www.movimed.com/Illumination.htm National Instruments 2013. Connect LabVIEW to any industrial Network and PLC. Accessed 13th January 2013. http://www.ni.com/white-paper/5407/en National Instruments 2012. Introduction to OPC. Accessed 7th March 2013. http://www.ni.com/white-paper/7451/en

Automation of DMPSManufacturing by using LabView & PLC (F. Fareeza)

332



ISSN: 2252-8814

[18] National Instruments 2012. LabVIEW 2012 Datalogging and Supervisory Control Module Help. Accessed 19th March 2013. http://digital.ni.com/manuals.nsf/websearch/51AA757C42C65F0D86257A0C0029ABD0 [19] National Instruments 2008. Robotics Fundamental Series: Feature Extraction. Accessed 18th April 2013. http://www.ni.com/white-paper/8222/en [20] National Instruments 2005. NI Vision for LabVIEW User Manual. Accessed 21st Febru- ary 2013. http://www.ni.com/pdf/manuals/371007b.pdf [21] National Instruments 2006. Counting Particles or Cells Using IMAQ. Accessed 19th March 2013. http://www.ni.com/white-paper/316 [22] National Instruments 2011. Geometric Matching Example. Accessed 22nd February 2013. http://zone.ni.com/devzone/cda/epd/p/id/5555 [23] National Instruments 2011. Bar Code Example. Accessed 23rd March 2013. http://zone.ni.com/devzone/cda/epd/p/id/5533 [24] Pokharel, B. 2013. HAMK University of Applied Sciences. Degree Programme in Au- tomation Engineering (20092013). [25] Quality Digest 2001. The Future of Machine Vision. Site Editor John E. Agapakis. Ac- cessed 7th January 2013 http://www.qualitydigest.com/oct98/html/machfutr.html [26] Sudcamp 2012. A Technology Hub. Accessed 23rd March 2013. http://www.sudcamp.com/ccd-vs-cmos-imagesensor-in-digital-camera/ [27] Vision Systems Oy. 2012. Robolapaper. Education Materials. [28] Wikimedia Commons 2012. File: Ccd-sensor.jpg. Accessed 18th March 2013. http://commons.wikimedia.org/wiki/File:Ccd-sensor.jpg [29] WiseGeek 2011. What is Machine Vision Image Processing. Site Editor Andrew Jones. Accessed 20th January 2013. [30] http://www.wisegeek.com/what-is-machine-vision-image-processing.htm#did-you- know

BIOGRAPHIES OF AUTHORS My name is Fareeza F I was born in 1983. I completed B.E. (Electronics & Instrumentation) from University of Madras, , Chennai in 2004. I have completed M.Tech. (Instrumentation & Control) from Bharath Institute of Higher Education & Research (Deemed University), Chennai. I have completed Ph.D. (Instrumentation) from Rayalaseema Univeristy in 2018. I have a total of 11 years at Dr. M.G.R. Educational & Research Institute, Chennai in Department of Electrical & Electronics Engineering.

My name is Chunchu Rambabu I was born in 1980. I completed B.Tech. (Electronics & Control) from Jawharlal Technical University, Hyderabad in 2004. I have completed M.Tech. (Instrumentation & Control) from Bharath Institute of Higher Education & Research (Deemed University), Chennai. I have completed Ph.D. (Instrumentation) from Rayalaseema Univeristy in 2018. I have a total of 11 years. Currently working as Lecturer in Arba Minch University, Sawla Campus

My name is S.Krishnaveni I had completed B.E. from the Department of Electrical and Electronics Engineering, Mepco Schlenk engineering college, Madurai Kamaraj University, Tamil Nadu, in 1996 and got my Master of Engineering degree First Class in Power Electronics and Industrial Drives, Sathyabama University, Tamil Nadu, India in June 2003. I have teaching experience of more than 19 years. For the past 13 years, Iam working as an Associate professor in EEE department in Dr M.G.R Educational and research institute university, Chennai, Tamil Nadu. I am a Professional member and advisor for 2 IEEE societies namely Industrial Application society and Women in Engineering.

IJAAS Vol. 7, No. 4, December 2018: 322 – 333

IJAAS

ISSN: 2252-8814



333

My name is Abel Chernet Kabiso I was born form my father ChernetKabiso and mother TagesechHadero in 1994, I graduated from Addis Ababa Science and Technology University in 2017/18 BSc Electromechanical Engineering and currently I’m working as Department Head of Electromechanical Engineering in Arba Minch University Sawla Campus.

Automation of DMPSManufacturing by using LabView & PLC (F. Fareeza)

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 334~346 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp334-346



334

Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus Images Using CBRIR Based on Lifting Wavelets S.S.Tadasare, S.S.Pawar Department of Electronics and Telecommunication Engineering, Bharati Vidyapeeth’s College of Engineering, Kolhapur, India

Article Info

ABSTRACT

Article history:

In this paper we present a lifting wavelet based CBRIR image retrieval system that uses color and texture as visual features to describe the content of a retinal fundus images. Our contribution is of three directions. First, we use lifting wavelets 9/7 for lossy and SPL5/3 for lossless to extract texture features from arbitrary shaped retinal fundus regions separated from an image to increase the system effectiveness. This process is performed offline before query processing, therefore to answer a query our system does not need to search the entire database images; instead just a number of similar class type patient images are required to be searched for image similarity. Third, to further increase the retrieval accuracy of our system, we combine the region based features extracted from image regions, with global features extracted from the whole image, which are texture using lifting wavelet and HSV color histograms. Our proposed system has the advantage of increasing the retrieval accuracy and decreasing the retrieval time. The experimental evaluation of the system is based on a db1 online retinal fundus color image database. From the experimental results, it is evident that our system performs significantly better accuracy as compared with traditional wavelet based systems. In our simulation analysis, we provide a comparison between retrieval results based on features extracted from the whole image using lossless 5/3 lifting wavelet and features extracted using lossless 9/7 lifting wavelet and using traditional wavelet. The results demonstrate that each type of feature is effective for a particular type of disease of retinal fundus images according to its semantic contents, and using lossless 5/3 lifting wavelet of them gives better retrieval results for almost all semantic classes and outperform 4-10% more accuracy than traditional wavelet.

Received Apr 23, 2018 Revised Jun 9, 2018 Accepted Jul 26, 2018 Keyword: Content Based Retinal Image Retrieval Haemorrhages Lifting Wavelet, Exudates Microaneurysms Retina

Corresponding Author: S.S.Tadasare, Department of Electronics and Telecommunication Engineering, Bharati Vidyapeeth’s College of Engineering, Kolhapur, India. Email: ss.tadasare@gmail.com

INTRODUCTION Diabetes has become one of the rapidly increasing health threats worldwide [21]. Only in Finland, there are 30 000 people diagnosed to the type 1 maturity onset diabetes in the young, and 200 000 people diagnosed to the type 2 latent autoimmune diabetes in adults [4]. In addition, the current estimate predicts that there are 50 000 undiagnosed patients [4]. Proper and early treatment of diabetes is cost effective since the implications of poor or late treatment are very expensive. In Finland, diabetes costs annually 505 million euros for the Finnish health care, and 90% of the care cost arises from treating the complications of diabetes [5]. These alarming facts promote the study of automatic diagnosis methods for screening over large Journal homepage: http://iaescore.com/online/index.php/IJAAS

IJAAS ISSN: 2252-8814  335 populations. Fundus imaging has an important role in diabetes monitoring since occurrences of retinal abnormalities are common and their consequences serious. However, since the eye fundus is sensitive to vascular diseases, fundus imaging is also considered as a candidate for non-invasive screening. The success of this type of screening approach depends on accurate fundus image capture, and especially on accurate and reliable image processing algorithms for detecting the abnormalities. Numerous algorithms have been proposed for fundus image analysis by many research groups [13, 6, 25, 15, and 18]. However, it is impossible to judge the accuracy and reliability of the approaches because there exists no commonly accepted and representative fundus image data base and evaluation protocol. With a widely accepted protocol, it would be possible to evaluate the maturity and state-of-the-art of the current methods, i.e., produce the achieved sensitivity and selectivity rates. For example, commonly accepted strict guidelines for the evaluation of biometric authentication methods, such as the FERET and BANCA protocols for face recognition methods [16, 2], have enabled the rapid progress in that field, and the same can be expected in medical image processing related to diabetic retinopathy detection. The main contribution of this work is to report a publicly available diabetic retinopathy database, containing the ground truth collected from several experts and a strict evaluation using proposed work of CBRIR. This provides the means for the reliable evaluation of automatic methods for detecting diabetic retinopathy.

DIABETIC RETINOPATHY In the quality 1 diabetes, the insulin concept in the pancreas is perpetually damaged, whereas in the quality 2 diabetes, the higher animal is abyss from increased intervention to insulin. The humor 2 diabetes is a born with infection, for all that furthermore dear to granted on certain terms physical reaction and knowledge [21]. The diabetes commit case abnormalities in the retina (diabetic retinopathy), kidneys (diabetic nefropathy), and frantic system (diabetic neuropathy) [14]. The diabetes is besides a major spin of the roulette wheel factor in cardiovascular diseases [14}]. The diabetic retinopathy is a microvascular difficult situation of diabetes, at the bottom of abnormalities in the retina, and in the worst situation, blindness. Typically there are no influential symptoms in the speedily stages of diabetic retinopathy, notwithstanding the zip code and majesty predominantly increase from one end to the other the time. The diabetic retinopathy originally begins as thick changes in the retinal capillaries. The sooner detectable abnormalities are mircroaneurysms (Ma) (Figure. 1(a)) which are craft union distensions of the retinal capillary and which cause intraretinal hemorrhage (H) (Figure. 1(b)) when ruptured. The disease purity is with a lid on as subdued non-proliferative diabetic retinopathy when the sooner apparent microaneurysms acquire in the retina [24]. In anticipate, the retinal edema and intimately exudates (He) (Figure. 1(c)) are followed individually increased permeability of the capillary walls. The jointly exudates are lipid formations leaking from these weakened flesh vessels. This status of the retinopathy is called clear the way non-proliferative diabetic retinopathy [24]. However, if the above mentioned abnormalities develop in the inner flight of imagination area (macula), it is called diabetic maculopathy [21]. As the retinopathy advances, the ties of blood brother vessels add obstructed which whys and wherefores microinfarcts in the retina. These microinfarcts are called peaceful exudates (Se) (Figure. 1(d)). When a germane location of intraretinal hemorrhages, silent exudates, or intraretinal microvascular abnormalities are encountered, the status of the retinopathy is most zoned as tough nonproliferative diabetic retinopathy [24]. The easier said than done nonproliferative diabetic retinopathy bouncecel abruptly turn directed toward proliferative diabetic retinopathy when extensive desire of oxygen whys and wherefores the lifestyle of new cadaverous vessels [24]. This is called as neovascularisation (Figure. 1(e)) which is a real glare sight intended state. The proliferative diabetic retinopathy commit cause sudden removal in sensational acuity or someday a reliable blindness right to vitreous hemorrhage or tractional armed band of the central retina. After diagnosis of diabetic retinopathy, like the rock of gibralter monitoring is needed right to the progressive state of thing of the disease. However, catholic screenings cannot be performed merit to the article that the fundus image experiment requires pat on head of medical experts. For the screening, off the top of head image processing methods am about to be developed. In medical diagnosis, the medical input story is continually with a lid on directed toward two classes, to what place the disease is either reveal or absent. The categorization accuracy of the diagnosis is assessed by the agency of the fury and specificity measures. Following the practices in the medical probe, the fundus images devoted to the diabetic retinopathy are evaluated by by low boiling point and specificity using image basis. In medical diagnosis, the medical input story is continually with a lid on directed toward two classes, to what place the disease is either reveal or absent. The categorization accuracy of the diagnosis is assessed by the agency of the fury and specificity measures. Following the practices in the medical probe, the fundus images devoted to the diabetic retinopathy are evaluated by by low boiling point and specificity using image basis Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus… (S.S.Tadasare)

336



ISSN: 2252-8814

Figure 1. Abnormal findings in the eye fundus caused by the diabetic retinopathy: (a) microaneuryms (marked with an arrow), (b) hemorrhages, (c) hard exudates, (d) soft exudate

Negative

True Negative (TN)

False Negative (FN)

(Marked with an arrow), and (e) neovascularization.

CURRENT EVALUATION PRACTICES Sensitivity is the percentage of abnormal fundus classified as abnormal, and specificity is the percentage of normal fundus classified as normal by the screening. The higher the sensitivity and specificity values, the better the diagnosis. Sensitivity and specificity can be computed as [22]:

Table 1. Perfomance Evaluation Test Result

Present

Absent

Positive

True Positive (TP)

False Positive (FP)

where TP is the number of abnormal fundus images found as abnormal, T N is the number of normal fundus images found as normal, FP is the number of normal fundus images found as abnormal (false positives) and FN is the number of abnormal fundus images found as normal (false negatives). Sensitivity and specificity are also referred to as the true positive rate (TPR) and true negative rate (TNR), respectively.

AUTOMATIC METHODS As mentioned previously, the diagnosis of diabetic retinopathy can be divided into the following two categories: 1. Screening of the diabetic retinopathy 2. Monitoring of the diabetic retinopathy Most automatic systems approach the detection directly using shape, color, and domain knowledge of diabetic retinopathy findings, but the abnormalities can also be found indirectly by detecting changes between two fundus images taken from the same eye in different time moment [11, 17]. The direct approach contributes to screening of the disease, where indirect approach contributes to both screening and monitoring of the diabetic retinopathy. Both approaches use roughly the following stages for finding abnormalities in fundus images: 1) image enhancement 2) candidate diabetic retinopathy finding detection 3) classification to correct diabetic retinopathy category (or hypothesis rejection). Some of the main features distinguishing between the different findings and normal fundus parts are the color and brightness. The same features have IJAAS Vol. 7, No. 4, December 2018: 334 – 346

IJAAS ISSN: 2252-8814  337 been verified also by ophthalmologists. Unsurprisingly these features dominate in the automatic methods, and therefore will be shortly reviewed in our brief surveys. Most of the automatic methods also detect normal fundus parts, such as optic disk, blood vessels, and macula. The automatic methods either use the vital domain information provided by the normal fundus parts or remove them due to their similar color and shape appearance with abnormal fundus findings. The detection of normal fundus parts is not considered in this study.

5.   



 

HARD AND SOFT EXUDATES Used normal retinal findings (vasculature, optic disk, fovea, and abnormal findings) to estimate the illumination component using iterative robust homographic surface fitting to compensate the nonuniform illumination in fundus images using GABOR WAVELETS function In detection of bright diabetic retinopathy areas from fundus images applied adaptive local contrast enhancement to sub-image areas using the local mean and standard deviation of intensities and adjusted the image brightness through gamma correction. Using color Auto Correlogram function Determined abnormal and normal finding areas using intensity properties for dynamic clustering. From the result abnormal areas, hard exudates were separated from soft exudates and drusen using intensity contrast information between abnormal areas and immediate background. The domain knowledge of retinal blood vessels were used to remove false artifacts using colormoments function Eliminated the vessels by applying morphological closing to the luminance component of the fundus image. From the result, within a sliding window local standard variation image was calculated and thresholded into coarse exudate areas. More accurate countours were acquired by thresholding difference between original image and morphologically reconstructed image.used yellowish color and sharp edges to distinguish hard exudates from the fundus images. The image pixels were classified into background and yellowish objects using minimum distance discrimination, where the countour pixels of extracted optic disk were used as background color reference and pixels inside the contour were used as yellowish object color reference. The segmented yellowish areas and their edge information extracted with Kirsch’s mask were combined to hard exudate areas using lifting wavelets function Located the bright abnormal regions in fundus images by applying color transform clustering in RGB color space. The result areas were classified to hard exudates, soft exudates, and normal findings using support vector machine using HSV transform function Searched the coarse hard exudate areas using query image features masks with smoothed histograms of each color band of the fundus image. The segmented areas were classified to exudate and non-exudate regions using CBRIR. Color, region size, orientation, mean and standard deviation of intensity, and texture were used as features.

EVALUATION DATABASE A necessary tool for reliable evaluations and comparisons of medical image processing algorithms is a database of dedicatedly selected high-quality medical images which are representatives of the problem. In addition, information about the medical findings, the ground truth, must accompany the image data. An accurate algorithm should take the images as input, and produce output which is consistent with the ground truth. In the evaluation, the consistency is measured, and algorithms can be compared based on these performance metrics. In the following, we describe the images and ground truth for the diabetic retinopathy database. 6.1. Fundus Images The database consists of 89 colour fundus images of which 84 contain at least mild nonproliferative signs (Ma) of the diabetic retinopathy (two examples shown in Figures. 2(b) and 2(c)), and 5 are considered as normal which do not contain any signs of the diabetic retinopathy according to all experts participated in the evaluation (an example shown in Figure. 2(a)). The images were taken in the Kuopio university hospital. The images were selected by the medical experts, but their distribution does not correspond to any typical population, i.e., the data is biased and no a priori information can be devised from it. The diabetic retinopathy abnormalities in the database are relatively small, but they appear near the macula which is considered to threaten the eyesight. Images were captured with the same 50 degree field-of-view digital fundus camera with varying imaging settings (flash intensity, shutter speed, aperture, and gain) controlled by the system. The images contain a varying amount of imaging noise, but the optical aberrations (dispersion, transverse and lateral chromatic, spherical, field curvature, coma, astigmatism, distortion) and photometric accuracy (colour or intensity) are the same. Therefore, the system urged on photometric variance around the audio auditory Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus… (S.S.Tadasare)

338



ISSN: 2252-8814

range of vision of the diverse retinopathy findings gave a pink slip be proposed as small. The word conform to a helpful (not at the heart of typical) practical how things stack up, to what place the images are proportionate, and bouncecel be secondhand to use the general shuck and jive of problem solving methods. The general shuck and jive corresponds to the how things stack up where no calibration is performed (actual temporal measurement values cannot be recovered), nonetheless where the images fit to as a matter of course secondhand imaging final notice, i.e., the demand encountered in hospitals. This data exist is specified as “calibration candidly 1 fundus images”. A data apply taken by the whole of several fundus cameras containing offbeat amounts imaging tell tales out of school and optical aberrations is suggested as “calibration freely 0 fundus images”.

Figure 2. Examples of DIARETDB1 fundus images: (a) normal fundus, (b) abnormal fundus, and (c) abnormal fundus after treatment by photocoagulation.

6.2. Ground Truth The practically having to do with veracity measures for medical diagnosis methods are tiff and specificity (see Sec. 4.2 for the definitions). Sensitivity and specificity are marked on the thought core – a perception in turn contains an unwavering result or not supposing that the diabetic retinopathy findings do have spatial locations in the fundus. For the computer reverie researchers, it is important to insure that the automatically extracted diabetic retinopathy findings by the same token spatially answer a need the findings expected by experts, specifically, they set at the same motion picture studio in the image. Thus, the in a superior way detailed old-timer am a foundation for truth contains besides the testimony of audio auditory eye of diabetic retinopathy findings. For individually fundus theory there is an exact bolster truth had the law on in database. 

Marking visual findings The theory groundtrtuh is based on expert-selected findings thick to the diabetic retinopathy and healthy fundus structures (see Figure. 3). A person by the whole of a medical advancement M.D.) and specialization to ophthalmology is proposed as an expert.  Data format The expert knowledge gathered by all of the hold truth power plant is brought together to a thought file. Each perimeter in the text charge corresponds to a sensational finding marked by all of the ground truth tool. The data format for visual finding is bounded as

Figure 3. Structural elements of a normal fundus.

IJAAS Vol. 7, No. 4, December 2018: 334 – 346

IJAAS

ISSN: 2252-8814



339

Figure 4. Graphical directives for marking the visual findings using CBIR 

Training and Test Sets The exist of 130 images was sovereign into 5 image categories, and a tense number of randomly occupied images were taken from each sector to art an element of the workout set. The surplus of the images formulate the confirm set. The image categories were formed to prove that each diabetic retinopathy finding name of tune is included in the both preparation and test sets. The diabetic retinopathy finding types that each image accumulation contains are the following: 1. Red Small drug, haemorrhages, strictly exudates. 2. Red Small drug, haemorrhages, jointly exudates, reticent exudates. 3. Red Small blotter acid, haemorrhages, intimately exudates, could hear a pin drop exudates, neovascularisation. 4. Red low dots, haemorrhages, could hear a pin drop exudates, neovascularisation. 5. Normal. The categories call a spade a spade the typical advance of the diabetic retinopathy [17]. 

Evaluation protocol Methods used for automatic detection of diabetic retinopathy are evaluated by using sensitivity and specificity per image basis. Sensitivity is the percentage of abnormal fund uses classified as abnormal by the screening method and specificity is the percentage of normal fundus classified as normal by the screening method. The higher the sensitivity and specificity values, the better the method. Sensitivity and specificity values are calculated for three diabetic retinopathy finding classes: exudates (soft and hard), haemorrhages and red small dots (microaneurysm). In the current database the neovascularisation is not included due to its rarity. Sensitivity and specificity can be computed as [20]:

LITERATURE REVIEW Content based image retrieval for general-purpose image databases is a highly challenging problem because of the large size of the database, the difficulty of understanding images, both by people and computers, the difficulty of formulating a query, and the issue of evaluating results properly. A number of general-purpose image search engines have been developed. In the commercial domain, QBIC [7] is one of the earliest systems. Recently, additional systems have been developed such as T.J. Watson [18], VIR [10], AMORE [19], and Bell Laboratory WALRUS [20]. In the academic domain, MIT Photobook [8, 21] is one of the earliest systems. Berkeley Blobworld [22], Columbia Visualseek and Webseek [9], Natra [23], and Stanford WBIIS [24] are some of the recent well known systems. The common ground for CBIR systems is to extract a signature for every image based on its pixel values and to define a rule for comparing images. The signature serves as an image representation in the “view” of a CBIR system. The components of the signature are called features. One advantage of a signature over the original pixel values is the significant compression of image representation. However, a more important reason for using the signature is to gain an improved correlation between image representation and semantics. Actually, the main task of designing a signature is to bridge the gap between image semantics and the pixel representation, that is, to create a better correlation with image semantics [11]. Existing general-purpose CBIR systems roughly fall into three categories depending on the approach to extract signatures: histogram, color layout, and region-based search. There are also systems that combine retrieval results from individual algorithms by a weighted sum matching metric [4], or other merging schemes [25]. After extracting signatures, the next step is to determine a comparison rule, including a querying scheme and the definition of a similarity measure between images. For most image retrieval systems, a query is specified by an image to be matched. We refer to this as global search since similarity is based on the overall properties of images. By contrast, Efficient Content Based Image Retrieval there are also “partial search” querying systems that retrieve results based on a particular region in an image [26]. Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus… (S.S.Tadasare)

340



ISSN: 2252-8814

7.1. Feature Based CBIR Systems Some of the existing CBIR systems extract features from the whole image not from certain regions in it; these features are referred to as Global features. Histogram search algorithms [7] characterize an image by its color distribution or histogram. Many distances have been used to define the similarity of two color histogram representations. Euclidean distance and its variations are the most commonly used. The drawback of a global histogram representation is that information about object location, shape and texture is discarded. Color histogram search is sensitive to intensity variations, color distortions, and cropping. The color layout approach attempts to overcome the drawback of histogram search. In simple color layout indexing [7], images are partitioned into blocks and the average color of each block is stored. Thus, the color layout is essentially a low resolution representation of the original image. A relatively recent system, WBIIS [24], uses significant Daubechies' wavelet coefficients instead of averaging. By adjusting block sizes or the levels of wavelet transforms, the coarseness of a color layout representation can be tuned. Hence, we can view a color layout representation as an opposite extreme of a histogram. At proper resolutions, the color layout representation naturally retains shape, location, and texture information. However, as with pixel representation, although information such as shape is preserved in the color layout representation, the retrieval system cannot perceive it directly. Color layout search is sensitive to shifting, cropping, scaling, and rotation because images are described by a set of local properties [4]. Image retrieval using color features often gives disappointing results, because in many cases, images with similar colors do not have similar content. This is due to the fact that global color features often fails to capture color distributions or textures within the image. D. Zhang [27] proposed a method combining both color and texture features to improve retrieval performance. By computing both the color and texture features from the images, the database images are indexed using both types of features. During the retrieval process, given a query image, images in the database are firstly ranked using color Efficient Content Based Image Retrieval features. Then, in a second step, a number of top ranked images are selected and re-ranked according to their texture features. Two alternatives are provided to the user, one is the retrieval based on color features, and the other is retrieval based on combined features. When the retrieval based on color fails, the user will use the other alternative which is the combined retrieval. Since the texture features are extracted globally from the image; they are not an accurate description of the image in some cases, which degrades the system performance.

PROPOSED CBRIR SYSTEM In our proposed CBRIR system, we use the same features of texture and color, as visual features to represent each region extracted from the retinal fundus images. 8.1. Texture Feature Extraction In the existing region based CBIR systems, visual features are extracted on each pixel that belongs to the region, and each region is described by the average value of these pixel features. However, we have found out that these average feature values are not efficient in describing the region’s content. Also, these features are extracted from each pixel or text on for the purpose of segmentation and differ with different segmentation algorithms. We propose to extract the color and texture features from each image region Efficient Content Based Image Retrieval as a whole after being extracted from the segmented image, this will help in representing the region efficiently and will make us free to use any image segmentation method without being obliged to use the same features used in that segmentation method. The lifting wavelets CDF9/7 is used for lossy and SPL5/3 for lossless transformation to extract the texture information of retinal fundus images.

IJAAS Vol. 7, No. 4, December 2018: 334 – 346

IJAAS

ISSN: 2252-8814



341

Figure 5. Block diagram, of proposed system CBRIR based on lifting wavelet for retinal fundus images

8.2. Color Feature Extraction We use the HSV color space for color feature extraction, since it is natural, perceptually uniform, and easy to be converted to RGB space and vice versa. As the image regions extracted from the image after segmentation are approximately color homogeneous, it is possible to use the average HSV value in each channel of all pixels in the region as its perceptual color. We also use the standard deviation for each color channel resulting in six color features. The Min-Max normalization formula is used to have the values of each color feature in the range [0, 1]. 8.3. Region Percentage Area The last feature we use is the region percentage area of an image. We propose that the area occupied by a region in an image gives information on the importance of this region and this importance should be great for regions with larger areas proportionally to the image area. 8.4. Region Matching An image region is described by a feature vector of 31 normalized attributes named as f1 to f31. The first 24 features are for texture, f25 to f30 are for color, and f31 for region percentage. To measure the similarity between two images we have to compare each region in one image to all the regions in the other, and this comparison is based on the extracted region features. We use the Euclidian distance between the feature vectors to match two regions such that as the distance increases the matching between the two region decreases and vice versa. The distance between two image regions Ri and Rj denoted by dij is defined as:

(1) Where fki and fkj are the kth feature of the regions Ri and Rj, respectively, and WT and Wc are weights for texture and color features. Experimentally in our simulation we examined some values for WT and Wc and we chosen to set WT = 1, and Wc = 2, since we have texture features, whereas the color and area features are only Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus… (S.S.Tadasare)

342



ISSN: 2252-8814

seven, and thus we have to increase the effect of the color features on the distance measure between image regions, the effect of changing the values of WT and Wc on the retrieval performance will be one of our future work. The distance dij between any two image regions will be used to measure the overall similarity between a query image and a database image. 8.5. Image Similarity Search Given the definition of the distance between two regions, we are ready to compute the global similarity between two images. Suppose that we have query image IQ with M regions and database image ID with N regions, we compute the global similarity between the two images IQ and ID using the following procedure: Step 1: Using Equation 1, compute the distance between one region in IQ and all regions in ID. For each region Ri in the query image IQ, the distance between this region and the database image ID is defined as: (2) Where dij is the distance between Ri and any region Rj in the database image. This definition takes the minimum distance between the query region Ri and all the regions in the database image ID, which maximizes the similarity between the region and the database image. Step 2: We compute the similarity between the query image IQ and the database image ID as follows:

(3) Where αt is the weight for region Ri in image IQ, we use the percentage of the region in an image f31 as its weight (i.e. αt = ft31), since we think a region with a larger area plays a more significant role in contributing to the overall similarity value between two images than a region with a smaller area. Step 3: The similarity distance between the query image and the database image given in Equation 3 is not symmetric, to make it symmetric we compute the distance between the database image and the query image by repeating steps 1 and 2 for the regions in the database image, we define the distance between region Rj in the database image and the query image as: (4) This definition takes the minimum distance between the database image region Rj and all the regions in the query image IQ, which maximizes the similarity between the region and the query image. Step 4: The distance between ID and IQ can be defined as:

(5) Where αj is the weight for region Rj in image ID, and also we use it as the fj31 just as for the query image regions. In Figure 2, a line from a query region to a DB region corresponds to the minimum distance from the region in image IQ (for example with 7 regions) to the region in database image ID (with 9 regions). Whereas, a line from a DB region to a query region corresponds to the minimum distance from the region in image ID to the region in IQ. These distances are then added and divided by two to get the symmetric distance between image IQ and IQ as in step 5.

IJAAS Vol. 7, No. 4, December 2018: 334 – 346

IJAAS

ISSN: 2252-8814



343

Step 5: The overall distance between the two images IQ and ID is defined as:

(6) This definition of the distance between two images captures the overall similarity measure based on regional and global matching. As compared with many existing similarity measures in the literature, this definition strives to incorporate as much Efficient Content Based Image Retrieval semantic information as possible, and at the same time also achieves a computational efficiency. Given this definition, for each query image IQ, it is straightforward to compute Dist (ID,IQ) for every image ID in the database in the retrieval process.

Figure 6. Minimum Distance of Regions from Image IQ to Image ID and Vice Versa.

8.6. Image Retrieval Methodology Data Insertion The image retrieval system we propose in this chapter first segments each image in the database into distinct regions considered as objects in that image using the TBES algorithm. Features are extracted from each image region using lifting wavelet; these features are stored in the database files. We implement clustering with self-organizing map algorithm in the database feature space to group those regions of similar visual features into separate clusters to reduce the searching time in the query process. The SOM is chosen to have two dimensional 10×10 nodes in grid top topological organization, each of these nodes is considered as a cluster center. Each image region in the database is given a cluster number stored with it at the end of SOM training using the region’s features.

Figure 7. GUI for Identification and classification of abnormalities of retinal using CBRIR based on lifting wavelets for fundus images Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus… (S.S.Tadasare)

344



ISSN: 2252-8814

Query Image Processing Given a query image, our system processes the query as follows: Step 1: Perform the query image segmentation to obtain all the regions, say we have N regions (Qi : i = 1to N) in the query image. Step 2: Calculate the closest SOM node, also known as the best matching unit (BMU), to the query image regions feature vector to determine which class Qi belongs to. Assume that region Qi belongs to class Cj. Step 3: Retrieve all the regions in the database that belong to the class Cj. These regions constitute a region set X. The images containing any regions in the set X is subsequently retrieved. These images comprise an image set T and are the candidate images. Step 4: Compare the query image with the images in the image set T. The distance Dist (Q, I) given in Equation 6.10 is used to measure the similarity between the query image and a candidate image, and the topleast-distance images are returned to the user.

RESULTS AND SYSTEM EVALUATION In this section we present an evaluation of the proposed CBRIR systems based on traditional wavelet and proposed lifting wavelets. We also compare their performance with traditional wavelet system. The database consists of 89 colour fundus images of which 84 contain at least mild non-proliferative signs (Microaneurysms) of the diabetic retinopathy, and 5 are considered as normal which do not contain any signs of the diabetic retinopathy according to all experts who participated in the evaluation. Images were captured using the same 50 degree field-of-view digital fundus camera with varying imaging settings. This algorithm achieves a true positive rate of 100% for, false positive rate of 2.12% for hemorrhages and accuracy score 100% for microaneurysms and 94-98% for others. Table 1 shows Performance Evaluation and table 2 shows 91% accuracy with traditional wavelet.

Table 2. Performance evaluation CBRIR based on Lifting wavelet for retina fundus images Sr. No.

Classification

Training

Tested

Sensitivity

Specificity

Accuracy

Hard Exudates

100.00

1.12

95.45

Soft Exudates

100.00

1.12

94.12

Microaneurysms

100.00

0.00

100.00

Hemorrhages

100.00

2.25

90.48

Normal

95.51

1.12

80.00

Total Fundus Images

95.51

1.14

98.88

10. DISCUSSION For this algorithm we have used Image processing techniques like CBIR based on lifting wavelet from RGB image because CBIR along with HSV have high intensity as compare to Red and Blue, then hard thresholding function for highlight the fundus image, lifting wavelet for enhancement for the complemented image, and for manipulating these techniques we have used MATLAB 2015a and with the help of this tool we have design one GUI for Content Based Retinal Image Retrieval using Lifting Wavelet Transform for classification and identification of abnormal retinas from Diaretdb1 retinal database. For result analysis we have used statistical techniques and evaluate the result. One of the main contribution of the proposed CBRIR based on lifting wavelet method is taking discrete thresholding correspond to abnormal fundus image. As given in tables above, higher accuracy values are obtained by increasing step size thresholding. This provides us to produce an automatic solution for a general purpose without any need to manually label retinal mask. The next significant feature of developed system is using unsupervised classification approach which provides us to retinal blood vessels without any training operation. Additionally, HSV followed by proposed system is a new combination of methods and relatively better than the others. Hard discrete thresholding scheme gave us better segmentation results than existing [1]. The result in Table1 and table2 ensures the difference between for 89 retrieved images responding to the selected queries. CBRIR based on lifting wavelet is more effective than CBRIR based on traditional wavelets for fundus retinal images.

IJAAS Vol. 7, No. 4, December 2018: 334 – 346

IJAAS ISSN: 2252-8814  345 11. CONCLUSION Content based image retrieval is a challenging method of capturing relevant images from a large storage space. Although this area has been explored for decades, no technique has achieved the accuracy of human visual perception in distinguishing images. Whatever the size and content of the image database is, a human being can easily recognize images of same category. From the very beginning of CBIR research, similarity computation between images used either region based or global based features. Global features extracted from an image are useful in presenting textured images that have no certain specific region of interest with respect to the user. In this paper, we presented a content based retinal image retrieval that classify depending on disease to answer an image query, which are to use either normal, abnormal patient based features of retinal fundus images. We use Lifting wavelets, which is a powerful texture extraction technique either in describing the content of image regions or the global content of an image. Color histogram along with HSV as a global color feature and histogram intersection as color similarity metric combined with lifting texture have been proved to give 98-94% accuracy as good retrieval results as that of traditional wavelets by 4-10%.

12. FUTURE WORK The following developments can be made in the future: 1. Region based retrieval systems are effective to some extent, but their performance is greatly affected by the segmentation process. Development of an improved image segmentation algorithm is one of our future works. 2. To further improve the performance of the retrieval system, the study of taking shape features into account during similarity distance computation can be considered. 3. To obtain better performance, the system can automatically pre-classify the database into different semantic images (such as cancer tissue, kidney stone, tumor tissue, texture vs. non texture images) and develop algorithms that are specific to a particular semantic image class. 4. Demonstration of using different color and texture weights in Equation 2 and their effect on the retrieval results.

ACKNOWLEDGEMENTS We are thankful to Bharati vidyapeeth for providing us a plateform for under taken post graduate project. Also we are thankful to S. S. Pawar for timely supervision and support.

REFERENCES [1]

[2] [3]

[4]

[5]

[6]

[7] [8]

Manjiri B. Patwari, Dr. Ramesh R. Manza,Yogesh M. Rajput, Dr. Manoj Saswade, Dr. Neha K. Deshpande, Sangramsing N. Kayte, “Calculation of Retinal Blood vessels Tortuosity by using Image processing Techniques and Statistical Techniques”, 2nd International Conference on System Modeling & Advancement in Research Trends (SMART) Department of Computer Applications, TMIMT ,Teerthanker Mahaveer University, 2013. Swati Agarwal, A. K. Verma, Nitin Dixit, “Content Based Image Retrieval using Color Edge Detection and Discrete Wavelet Transform”, IEEE, 978-1-4799-2900-9/2014. Swati Agarwal, A. K. Verma, Preetvanti Singh, "Content Based Image Retrieval Using Discrete Wavelet Transform & Edge Histogram Descriptor", IEEE International Conference on Information System and Computer Networks (ISCON), pp. 19-23, March 2013. Balamurugan, P. Anandhakumar, "An Integrated Color and Texture Feature based Framework for Content Based Image Retrieval using 2D Wavelet Transform", IEEE Proceedings of the International Conference on Computing, Communication and Networking, 2008. Seema Haribhau Jadhav, Shah Aqeel Aehmad, "Content Based Image Retrieval System with Hybrid Feaure set and Recently Retrieved Image Library", (UCSIS) International Journal of Computer Science and Information Security, Vol. 59, No.5, 2012. A. Mumtaz, S. A. M. Gilani, T. Jameel, "A Novel Texture Image Retrieval System based on Dual Tree Complex Wavelet Transform and Support Vector Machines", IEEE 2ad International Conference on Emerging Tecbnologies, 2006. S. Soman, M. Ghorpade, V. Sonone, S. Chavan, "Content Based Image Retrieval using Advanced Color and Texture Features", International Conference in Computational Intelligence (ICCIA), 2011. G. Quellec, M. Lamard, G. Cazuguel, B. Cochener, C. Roux, "Adaptive Non-Separable Wavelet Transfrm via Lifing ad its Application to Content-Base Image Retrieval", IEEE Trans Image Process, VoI.19, No.l, pp.25- 35, 2010.

Early Detection of High Blood Pressure and Diabetic Retinopathy on Retinal Fundus… (S.S.Tadasare)

346 [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

[21] [22]

[23] [24]

[25]

[26]

[27]

[28]

[29]



ISSN: 2252-8814

A. Komali, V. Satish Kumar, K. Ganapathi Babu, A. S. K. Ratnam, "3D Color Feature Extraction in Content-Based Image Retrieval", International Journal of Soft Computing and Engineering (IJSCE) Vol. 2, Issue 3, July 2012. Tian Yumin, Mei Lixia, "Image Retrieval Based on Multiple Features Using Wavelet", 5th IEEE International Conference on Computational Intelligence and Multimedia Applications (ICCIMA'03), 2003. B. Ramamurthy, K. R. Chandran, "Content Based Image Retrieval for Medical Images Using canny Edge Detection Algorithm", International Journal of Computer Applications (0975-8887), Vol. 17, No.6, Mar. 20 12. V.N. Gudivada and V.V. Raghavan, “Content based image retrieval systems”, IEEE Computer, 28 (9), 18-22, 1995. Xin Zhang and Guoliang Fan, “Retinal Spot Lesion Detection Using Adaptive Multiscale Morphological Processing”, Springer-Verlag Berlin Heidelberg 2006, ISVC 2006, LNCS 4292, pp. 490–501, 2006. Saiprasad Ravishankar, et al, “Automated Feature Extraction for Early Detection of Diabetic Retinopathy in Fundus Images”, 978-1-4244-3991- 1, 2009 IEEE. 2012. Keith A. Goatman, et al, “Detection of New Vessels on the Optic Disc Using Retinal Photographs”, IEEE transactions on medical imaging, vol. 30, no. 4, April 2011. B´alint Antal, et al, “An Ensemble-Based System for Microaneurysm Detection and Diabetic Retinopathy Grading”, IEEE transactions on biomedical engineering, vol. 59, no. 6, june 2012. Anderson Rocha, “Points of Interest and Visual Dictionaries for Automatic Retinal Lesion Detection”, IEEE transactions on biomedical engineering, vol. 59, no. 8, august 2012. Parisut Jitpakdee, et al, “A Survey on Hemorrhage Detection in Diabetic Retinopathy Retinal Images”, IEEE 2012, 978-1-4673-2025-2. Arti Yerolkar, Swati Madhe, “Blood Vessel Segmentation and Classification of Retinal Image for Detection of Proliferative Diabetic Retinopathy”. Jyoti D. Patil, Anant. L. Chaudhari, “Tool for the Detection of Diabetic Retinopathy using Image Enhancement Method in DIP”, International Journal of Applied Information Systems (IJAIS), Volume3, No3. 2012 – ISSN: 2249-0868. Knudtson MD, Klein BEK, Klein R, Wong TY, Hubbard LD, et al. Variation associated with measurement of retinal vessel diameters at different points in the pulse cycle. Br J Ophthalmol.2004; 88:57–61. S. Jiméneza,P. Alemanya, I. Fondónb, A. Foncubiertab, B. Achab and C. Serranob “Automatic detection of vessels in color fundus images” © 2009 Sociedad Española de Oftalmología. Published by Elsevier España, s.larchsocespoftalmol. 2010; 85(3):103-109. Ana MM, Aurelio C: Segmentation of Retinal Blood Vessels by Combining the Detection of Centerlines and Morphological Reconstruction. IEEE Trans Medical imaging 2006, 25:1200-1213. Fischer JG, Mewes H, Hopp HH, Schubert R. “Analysis of pressurized resistance vessel diameter changes with a low cost digital image processing device”. Comput Meth Prog Bio. 1996; 50:23–30. 2nd International Conference on System Modeling & Advancement in Research Trends (SMART) Department of Computer Applications, TMIMT ,Teerthanker Mahaveer University [2013]. Manjiri B. Patwari, Dr. Ramesh R. Manza, Dr. Manoj Saswade and Dr. Neha Deshpande, “A Critical Review of Expert Systems for Detection and Diagnosis of Diabetic Retinopathy”, Ciit International Journal of Fuzzy Systems, February 2012, DOI: FS022012001 ISSN 0974-9721, 0974-9608. Yogesh M. Rajput, Ramesh R. Manza, Manjiri B. Patwari, Neha Deshpande, “Retinal Blood Vessels Extraction Using 2D Median Filter”, Third National Conference on Advances in Computing (NCAC-2013), 5th to 6th March 2013, School of Computer Sciences, North Maharashtra University, Jalgaon-425001 (MS) India. Yogesh M. Rajput, Ramesh R. Manza, Manjiri B. Patwari, Neha Deshpande, “Retinal Optic Disc Detection Using Speed Up Robust Features”, National Conference on Computer & Management Science [CMS-13], April 25-26, 2013, Radhai Mahavidyalaya, Auarngabad-431003(MS India). Manjiri B. Patwari, Ramesh R. Manza, Yogesh M. Rajput, Manoj Saswade, Neha K. Deshpande, “Review on Detection and Classification of Diabetic Retinopathy Lesions Using Image Processing Techniques”, International Journal of Engineering Research & Technology (IJERT), ISSN: 2278-0181, Vol. 2 Issue 10, October - 2013, Impact Factor 1.76. Manjiri B. Patwari, Ramesh R. Manza, Yogesh M. Rajput, Neha K. Deshpande, Manoj Saswade, “Extraction of the Retinal Blood Vessels and Detection of the Bifurcation Points”, International Journal in Computer Application (IJCA), September 18, 2013. ISBN: 973-93-80877-61-7, Impact Factor 0.821.

IJAAS Vol. 7, No. 4, December 2018: 334 – 346

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 347~352 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp347-352



347

Recycling of Industrial Waste Water for the Generation of Electricity by Regulating the Flow Control Sensor using IoT Kaviyaraj.R, Karthika M.S, Jeni Narayanan L.A, Saleekha Sri Krishna College of Engineering and Technology, Kuniamuthur, Coimbatore, Tamil Nadu 641008, India

Article Info

ABSTRACT

Article history:

The paper focuses on generating the renewable energy source from industrial waste water effluents. Utilizing the industrial waste water in order to generate electricity, a flow control sensor has been installed at the outlet of the tunnel which passes the waste water to the turbine. As per the need, the generation of electricity varies with respect to the flow through the use of flow control sensor. The generated electricity is then used for powering the street lights, gardening and run-way paths, during night time. The flow control sensor when integrated using IoT and cloud storage facilitates efficiency and scalability thereby providing massive utilization of energy usage.

Received Apr 19, 2018 Revised Jun 28, 2018 Accepted Aug 1, 2018 Keyword: Flow Control Sensor Industrial Waste Water Internet Of Things (IoT) Micro Hydro Power Plant Industry Renewable Energy Turbine

Corresponding Author: Kaviyaraj.R,

Sri Krishna College of Engineering and Technology, Kuniamuthur, Coimbatore, Tamil Nadu 641008, India. Email: 17epcs009@skcet.ac.in

INTRODUCTION Now a days the waste water discharges from industries is one of the important pollution source in the pollution of the water environment. During the last years a huge amount of industrial waste water was discharged into rivers, lakes, and coastal areas. To treat these industrial waste water, there are different techniques available. But we can make use of this industrial waste water directly in order to generate electricity. This is a new shoot on this branch of energy production, which is capable of generating electrical energy from waste water. To make this in realistic way we have come into a smart water management model. This is totally managed by the sensors and the controllers without any human interference. The IoT sensors can monitor the water level in the tank and thereby control the flow of water. A network of smart meters that collect granular real-time data across the water tank that could help to identify the water leakage, hidden patterns in water consumption, use predictive analytics to regulate demand and supply and set up alarms for notifying anomalies. Then the water resource is proceeding for electricity generation. There are several techniques for treating waste water from industries but there is no efficient method in order to generate electricity from it. This paper introduces a new algorithm to collect and maintain industrial waste water for energy production.

LITERATURE SURVEY Nan Win Aung et.al [1] proposed a solution for solving the problem of fluctuation in frequency and voltage output from a generator due to variation in consumer load. The fluctuation in frequency and voltage will affect various electrical appliances adversely. The main reason for this problem is variation in speed of water turbine due to change of consumer load. The solution uses a microcontroller and liquid crystal display (LCD) to monitor the voltage, power, current and frequency of electronic load controller (ELC). ELC is a solid Journal homepage: http://iaescore.com/online/index.php/IJAAS

348



ISSN: 2252-8814

state electronic device which is used to regulate output power of a micro-hydro power system and there by handing load on the turbine. Asanka et.al [2] proposed a microcontroller based electronic governor and control system for mini hydro power plant, which is a complete requirement in the field of small scale power plants. This is to improve the features of mechanical governor systems and remote controllability in user friendly manner. Here it uses microcontroller technology with the electronic governor system, so it is cost effective. Mohamed Nazih Abdallh et.al [3] says about the best techniques for industrial waste water treatment of food industry. The industrial waste water will contain toxic materials and possibly pathogenic bacteria. So before discharging these waste water in to surface or ground water it should be fully treated. Compared to other industries the food industry require large amount of water. The waste water from food industry causes pollution problems due to its high COD (Chemical Oxygen Demand) and BOD (Biochemical Oxygen Demand).There are mainly two treatment technique which are physical-chemical treatment and physical-chemical treatment followed by biological treatment technique. Physical-chemical treatment can remove max degree of chemical oxygen demand, biochemical oxygen demand and suspended solids from food waste water. Anaerobic digestion is one of the suitable technique for the treatment of high strength organic effluents. Edumund et.al [4] developed a sensible and affordable system to generate a significant quantity of electric power using closed loop water power plant system. Here the proposed concept is generating electric power with the little amount of water. Hence it requires only small quantity of water, the wastage of water is minimized. The construction is very simple because it is a small power plant. It consists of two tanks with the help of external pressure. This is a fully automated system. Mu Han.et.al [5] proposes automatic flow control technique for fluids from different industries. For the frequency input of the micro controller a sensing unit, photo interrupter and slotted disk are using. Automatic control system is method which makes the needed adjustments without human aid. Balaji et.al [6] proposed a microcontroller based power generation from the agricultural field. That is to make agriculture simple, effective and automatic irrigation system in order to provide Water to the field based on their soil types. Here the moisture content of the soil will test using moisture sensor and the output is taken as the input of the microcontroller. By this result the microcontroller can control the motor, which is used pump the water from the ground and it will be stored in a tank. When the water reaches the tank outlet the turbine will rotate and the motion of the turbine helps to generate the power which is done based on the hydro electric power generation principle. The generated current can be used for the home applications and this system will preserve an optimum water distribution. Jinhuang Huang et.al [7] says about the intelligent water dispenser systems based on embedded systems. Heating and heat insulation are the main functions of current general domestic water dispenser systems. It is achieved by using the analog systems. But they have so many disadvantages such as it will affect by external environment factors. Microcontroller-based temperature monitoring and controlling is used in actual production. The MCV-based smart temperature control water dispenser system is used for the good functioning. Hydro turbine governing system [8] in which the main controller is the Governor of the hydraulic turbine and it varies the water flow in order to control the speed or power output. The size, type and cost vary according to their capacity and performance. Here speed sensing methods are used based on speed response. The speed governor is used to maintain the water turbine speed in spite of the variation of the load. It must be high sensitivity and stability. The waste water discharged from different industries has different characteristics and volume. Different aerobic or anaerobic biological processes are applied for the treatment of industrial waste water from food industries [9]. For the selection of treatment process the site conditions of the wastewater treatment plant, wastewater properties and economical efficiency of the treatment are considered.

IJAAS Vol. 7, No. 4, December 2018: 347 – 352

IJAAS ISSN: 2252-8814 3. METHODOLOGY 3.1. System Architecture Figure 1. Shown process the system Architecture :



349

Figure 1. System Architecture

3.2. COMPONENTS Arduino Arduino is open source PC equipment and programming organization which will plan and fabricates singleboard microcontrollers and microcontroller packs for building computerized gadgets that can sense and control objects in the physical and advanced world. Here it is used to control the working of flow control sensor, display, street lights. It connects with the internet and it can be access from anywhere. Battery The battery used here isLithium-Ion battery, the current generated is being stored in this battery. The amount of current in the battery will be shown on a display that is connected to the Arduino board, the stored current so obtained can be used to power up the street lights. Cloud The user input to the Arduino micro-controller could be given from anywhere using the cloud infrastructure. Display The Battery voltage will be displayed on the LED display. Also, the user inputs/commands will also be displayed. Flow control sensor Water flow sensor is used to control the flow of water and send it to the turbine for producing the current according to the water flow. Based on the input from the user/arduino the flow of water can be controlled. Generator It is used to generate the electricity from the turbine, it is done by converting the mechanical energy to electrical energy.

Figure 2. Generator Recycling of Industrial Waste Water for the Generation of Electricity by Regulating the ... (Kaviyaraj.R)

350



ISSN: 2252-8814

Step Up Transformer Transformer is an electrical device that transfers electrical energy between two or more circuits through the technique electromagnetic induction. In step up transformer the main task is converting high voltage and low current from the primary side to the low voltage and high current on the secondary side and vice versa.

Figure 3. Step Up Transformer

Street light The lights in the industry are powered up with the electricity produced from the generator with the help of turbine. Turbine Turbine is the main part used to produce the electricity from the flow of water. It is a machine which works by the action of fluid and it contains a circular set of blade. The moving fluid acts on the blades which spins turbine connected to the generator to produce electricity.

4. METHODOLOGY Step-1: Start and connect a microcontroller to the internet via wifi. Step-2: Collect the industrial waste water in the tank. Step-3: Based on the output of the flow control sensor it will work. Step-4: The water will flow into the turbine and it starts spinning. Step-5: The rotation of the turbine will produce electricity in the generator. Step-6: The step up transformer is used to amplify the generated energy. Step-7: The electrical energy is stored in lithium-ion battery and the voltage can be viewed in display. Step-8: The energy stored in the battery is used to power up the street light. Let the catchment basin be represented as δ. The water flow in the catchment basin varies with respect to time which can be defined as d (δ). The energy produced or the current generated is directly proportional to the kinetic energy, potential energy and the flow inside and outside the catchment basin. Therefore the level of current produced can be defined as 𝑦1 𝑡

+ 𝑡𝑣1 +

1 2

𝑔12 =

𝑦2 𝑡

+ 𝑡𝑣2 +

1 2

𝑔22

(1)

Since the potential, kinetic and the mechanical energy varies with respect to time which remains constant n both the outside of the catchment basin as well as the inside of the basin, the output will be constant. 𝑦 𝑡

+ 𝑡𝑣 +

1 2

𝑔2 = 𝑘

(2)

Since the water level varies with respect to time and rate of flow of waste water, this change in fluctuation can be defined as 𝑑(𝑦) 𝑡

+ 𝑉𝑡𝑣 + 𝑔𝑑𝑧 = 0

(3)

Since the tank is open to the atmosphere 𝑦1 = 𝑦2 IJAAS Vol. 7, No. 4, December 2018: 347 – 352

(4)

IJAAS

ISSN: 2252-8814



351

The rate at which the water comes into the catchment basin will be presumed as zero 𝑣1 = 0

(5)

Therefore the energy produced varies with respect to the flow or rate of flow of water and the moment at which the power can be generated are as follows: 𝑣2 = √2𝑦(𝑧1 − 𝑧2) = √2𝑔𝑚 𝑣22 −𝑣12 2𝑔

m=+pAV

(6)

The energy production also changes with respect to the kinetic energy when it moves inside the pump to the turbine which can be defined as 𝑃1 + 𝛾𝑤 𝑧𝑜 + 𝛾𝑚 𝑧 = 𝑃2 + 𝛾𝑤 𝑧𝑜 +γ𝑚 𝑧

(7)

𝑃2 − 𝑃1 = (𝛾𝑚 − 𝛾𝑤 )𝑧 𝑃1 −𝑃2

hA =

𝛾𝑤

+ 𝑣22 −

𝑣12 𝑔

Therefore the power generation can be defined as Power P=𝛾𝑤 𝑄ℎA

(8)

The efficiency or the current produced can be given as Efficiency =

𝑃 𝑃𝑢𝑚𝑝𝑜𝑢𝑡

(9)

CONCLUSION Since we are running out of renewable energy resources, it’s high time that we must go for methods to preserve them. The idea proposed here would produce electricity from industrial waste water which could be then used to light up streets. Innovative technologies could be used to extend this idea further for large scale electricity generation. Further, if turbine can be powered up, the electricity generation can be increased in large a quantity which facilitates powering up the street lights, etc.

REFERENCES [1] [2] [3]

[4]

[5]

[6]

[7] [8] [9]

Microcontroller Based Electrical Parameter Monitoring System of Electronic Load Controller Used in Micro Hydro Power Plant, Journal of Electrical and Electronic Engineering Volume 3, Issue 5, October 2015, Pages: 97-109. LPG Asanka, DMH Colombage, MHRH de Silva, SHN Tharanga “Microcontroller based Electronic Governor and Control System of a Mini-hydro Power Plant”Department of ELectrical Engineering, University of Moratuwa. Mohamed Nazih Abdallh, Walid Sayed Abdelhalim, Hisham Sayed Abdelhalim “Industrial Wastewater Treatment of Food Industry Using Best Techniques” International Journal of Engineering Science Invention ISSN (Online): 2319 – 6734, ISSN (Print): 2319 – 6726, www.ijesi.org ||Volume 5 Issue 8|| August 2016 || PP.15-28. M. F. Edmund laal, K. Mahesh kumar, M. Suresh, G. Saravanan, G. Deenadayalan. “Closed Loop Water Power Plant” International Journal Of Scientific & Technology Research Volume 2, Issue 10, October 2013 ISSN 22778616. Thwe Mu Han, Ohn Mar Myaing “Design and Construction of Microcontroller-Based Water Flow Control System” 2011 International Conference on Circuits, System and Simulation IPCSIT vol.7 (2011) © (2011) IACSIT Press, Singapore. Balaji S., DhivyaPriya E.L., Durairaj G., Illakkiya V. “Microcontroller Based Power Generation from the Agricultural Field” International Journal of Advanced Research in Electronics and Communication Engineering (IJARECE) Volume 4, Issue 4, April 2015. Huang, Jinhuang, and Jun Xie. "Intelligent water dispenser system based on embedded systems. "Mechatronics and Embedded Systems and Applications (MESA), 2010 IEEE/ASME International Conference on. IEEE, 2010. http://ahec.org.in/publ/Modern_Hydroelectric_Engineering_Practice_Prof_OD_Thapar/Volume_I/Chapter6_Hydro-Turbine_Governing_System.pdf. https://www.env.go.jp/earth/coop/coop/document/male2_e/007.pdf.

Recycling of Industrial Waste Water for the Generation of Electricity by Regulating the ... (Kaviyaraj.R)

352



ISSN: 2252-8814

[10] https://www.biologydiscussion.com/waste-management/waste-water-treatment/processes-of-waste-watertreatment-4-process-with-diagram/10989 [11] http://www.world-nuclear.org/nuclear-basics/electricity-generation-what-are-the-options.aspx [12] https://en.wikipedia.org/wiki/Electricity_generation

IJAAS Vol. 7, No. 4, December 2018: 347 – 352

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 353~360 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp353-360



353

Coding Schemes for Implementation of Fault Tolerant Parrallel Filter Nutenki Siddhartha, G Renuka Department of ECE, SR Engineering College, Hasanparthy, Warangal, Telangana, India.

Article Info

ABSTRACT

Article history:

Digital filters are utilized as a one of flag handling and correspondence frameworks. At times, the unwavering quality of those frameworks is basic, and blame tolerant channel executions are needed. Throughout the years, numerous systems that endeavor the channels' structure and properties to accomplish adaptation to internal failure have been proposed. As innovation scales, it empowers more unpredictable frameworks that join many channels. In those perplexing frameworks, it is regular that a portion of the channels work in parallel. A plan in view of big rectification coding has been as of late proposed to protect parallel channels. In that plan, each channel is deal with as a bit, and excess channels that go about as equality check bits are acquainted with distinguish and rectify blunders. In this short, applying coding systems to secure parallel channels is tended to in a broader manner. This decreases the assurance overhead and makes the quantity of excess channels autonomous of the quantity of parallel channels. The proposed technique is first described and then illustrated with two case studies. Finally, both the effectiveness in protecting against errors and the cost are evaluated for a field-programmable gate array implementation.

Received Mei 28, 2018 Revised Jul 10, 2018 Accepted Aug 18, 2018 Keyword: Error Correction Codes (ECCs) Filters Soft Errors

Corresponding Author: Nutenki Siddhartha, Department of ECE, SR Engineering College, Hasanparthy, Warangal, Telangana, India. Email: siddharthanutenki@gmail.com

INTRODUCTION Electronic circuits are progressively present in and space applications where dependability is basic. In those applications, the circuits need to give some level of adaptation to non-critical failure. This need is additionally expanded by the inherent unwavering quality difficulties of cutting edge CMOS innovations that incorporate, e.g., producing varieties and delicate blunders. Various strategies can be utilized to shield a circuit from blunders. Those range from adjustments in the assembling procedure of the circuits to decrease the quantity of mistakes to including repetition at the rationale or framework level to guarantee that blunders don't influence the framework usefulness. Channels are regularly utilized as a part of electronic frameworks to underline motions in certain recurrence ranges and reject motions in other recurrence ranges. In circuit hypothesis, a channel is an electrical system that adjusts the abundancy as well as stage qualities of a flag regarding recurrence. In a perfect world, a channel won't add new frequencies to the info flag, nor will it change the segment frequencies of that flag, yet it will change the relative amplitudes of the different recurrence segments as well as their stage connections. Today channels are generally utilized as a part of number of uses which in view of car, restorative, and space where unwavering quality of segments in computerized electronic circuits is basic. Channels or the like are basic in the operation of most electronic circuits. There are various bases of characterizing channels and these cover in a wide range of ways; there is no straightforward progressive grouping. As the behavioral properties

Journal homepage: http://iaescore.com/online/index.php/IJAAS

354



ISSN: 2252-8814

of flag changes the strategies of separating it will be contrast. Being particular with channel, the computerized channels have huge applications in advanced flag preparing. Sifting is likewise a class of flag preparing, the characterizing highlight of channels being the total or incomplete concealment of some part of the flag. It is in this manner in light of a legitimate concern for anybody associated with electronic circuit configuration to be able to create channel circuits fit for meeting a given arrangement of particulars. In flag handling, an advanced channel is a gadget or process that expels some undesirable segment or highlight from a flag. Advanced channels are utilized for two general purposes; division of signs that have been joined, and reclamation of signs that have been contorted somehow. Frequently, this implies expelling a few frequencies and not others with a specific end goal to smother meddling signs and decrease foundation commotion. Parallel Processing and correspondence frameworks. As a rule, Parallel channels are normally found in current flag the channels play out an indistinguishable preparing on various approaching signs from there is an inclination to utilize different input–multiple yield frameworks. The security of computerized channels has been broadly contemplated. For instance, blame tolerant executions in light of the utilization of deposit number frameworks or number-crunching codes have been proposed. The utilization of decreased accuracy replication or word-level insurance has been likewise examined another alternative to perform blunder remedy is to utilize two distinctive channel usage in parallel. Every one of those systems concentrate on the security of a solitary channel. This concise examinations the assurance of parallel channels utilizing more broad coding strategies. Specifically, a key contrast with ECCs is that both channel sources of info and yields are numbers. Hence, not just a zero or a one can be utilized for the coding (as finished with ECCs). This can be abused, as appeared in whatever is left of this brief, to give mistake remedy by including just two repetitive channels paying little heed to the quantity of parallel channels. The decreased number of repetitive channels does not influence the capacity of the plan to revise mistakes however diminishes the usage cost. In whatever is left of this short, to begin with, the parallel channels and the current ECC-based security plot are depicted. At that point, the proposed coding plan is given and delineated a couple of useful contextual analyses. At last, the contextual analyses are assessed for a field-programmable door cluster (FPGA) usage and contrasted and the already proposed ECC-based system. 1.1. Concept of Fault Tolerance Various procedures can be utilized to shield a circuit from mistakes. Those range from changes in the assembling procedure of the circuits to lessen the quantity of blunders to including repetition at the rationale or framework level to guarantee that mistakes don't influence the framework usefulness. Computerized Filters are a standout amongst the most generally utilized flag handling circuits and a few procedures have been proposed to shield them from blunders. There are number of techniques used to recognize flaws and the activities important to adjust the issues inside circuit. Advanced channels are broadly utilized as a part of flag handling and correspondence frameworks. There are distinctive adaptation to internal failure ways to deal with customary computational circuits and the DSP circuits. Now and again, the unwavering quality of those frameworks is basic, and blame tolerant channel executions are required. Throughout the years, numerous methods that endeavor the channels structure and properties to accomplish adaptation

LITERATURE REVIEW [1] In this paper, adaptation to non-critical failure construct framework based with respect to Error Correction Codes (ECCs) utilizing VHDL is outlined, executed, and tried. It recommends that with the assistance of ECCs i.e. Blunder Correction Codes there will be more ensured Parallel channel circuit has been conceivable. The channel they have utilized for blunder location and revision are for the most part limited motivation reaction (FIR) channels. They have been utilized Hamming Codes for blame adjustment in which they takes a square of k bits and produces a piece of n bits by including n−k equality check bits. The equality check bits are XOR mixes of the k information bits. By appropriately planning those mixes it is conceivable to distinguish and revise blunders. In this plan they have utilized repetitive module in which the information and equality check bits are store d and can be recuperated later regardless of the possibility that there is a blunder in one of the bits. This is finished by re - registering the equality check bits and contrasting the outcomes and the qualities put away. Along these lines utilizing hamming codes mistake can be recognized and revised inside the circuit. [2] In this paper, Triple Modular Redundancy (TMR) and Hamming Codes have been utilized to ensure distinctive circuits against Single Event Upsets (SEUs). In this paper, the utilization of a Novel Hamming approach on FIR Filters is considered and executed with a specific end goal to give low unpredictability, lessen deferral and territory proficient assurance methods for higher bits information. A novel Hamming code is proposed in this paper, to build the proficiency of higher information bits. In this paper, they IJAAS Vol. 7, No. 4, December 2018: 353 – 360

IJAAS ISSN: 2252-8814  355 have proposed method used to illustrate, how the part of overhead because of mixing the repetition bits, their consequent evacuation, cushion to cushion postpone in the decoder and utilization of aggregate range of FIR channel for higher bits are decreased. These depend on the novel hamming code usage in the FIR channel rather than regular hamming code used to ensure FIR channel. In this plan Hamming code utilized for transmission of 7-bit information thing. [3] In this paper, the outline of a FIR channel with self checking capacities in view of the deposit checking is broke down. Typically the arrangement of deposits used to check the consistency of the aftereffects of the FIR channel are based of theoretic contemplations about the dynamic range accessible with a picked set of buildups, the math qualities of the blunders caused by a blame and on the normal for the channel execution. This examination is frequently hard to perform and to acquire adequate blame scope the arrangement of picked buildups is overestimated. Gotten result and consequently requires that Instead, in this paper they have demonstrated how utilizing a comprehensive blame infusion battles permits to proficiently choose the best arrangement of buildups. Test comes about originating from blame infusion crusades on a 16 taps FIR channel showed that by watching the happened mistakes and the identification modules relating to various deposit has been conceivable to decrease the quantity of discovery module, while paying a little lessening of the level of SEUs that can be distinguished. Paired rationale overwhelms the equipment execution of DSP frameworks [4] In this paper they have proposed engineering for the execution of blame - tolerant calculation inside a high throughput multirate equalizer for a deviated remote LAN. The zone overhead is limited by abusing the logarithmic structure of the Modulus Replication Residue Number System (MRRNS). They had exhibited that for our framework the zone cost to amend a blame in a solitary computational channel is 82.7%. Adaptation to non-critical failure inside MRRNS design is executed through the expansion of excess channels. This paper has displayed a point by point investigation of the cost of executing single blame adjustment ability in a FIR channel utilizing the MRRNS. The blame tolerant design makes utilization of the mathematical properties of the MRRNS, and has been appeared to give critical range funds when contrasted and general systems. This engineering likewise requires couple of extra parts to be outlined, as indistinguishable excess channels are utilized, and the polynomial mapping stages are just extended from the first segments. Main Objectives of proposed method: There are various objectives over the protection of digital signal processing circuits. The main objectives are given as follows1). To achieve fault free digital circuit. 2). To detect and correct errors in digital circuit with more accuracy. 3). To reduce the overhead, needed for protection from error. 4). To improve efficiency. 5). To develop the application area from lower to higher order application.

ECC-BASED PROTECTION OF PARALLEL FILTERS The impulse response h[n] completely defines a discrete time filter that performs the following operation on the incoming signal x[n]:

(1) This property can be exploited in the case of parallel filters that operate on different incoming signals, as shown on Figure. 1. In this case, four filters with the same response process the incoming signals x 1[n], x 2[n], x 3[n], and x 4[n] to produce four outputs y 1[n], y 2[n], y 3[n], and y 4[n].

Figure 1. Block diagram for Parallel FIR Filters Coding Schemes for Implementation of Fault Tolerant Parrallel Filter (Nutenki Siddhartha)

356



ISSN: 2252-8814

To detect and correct errors, each filter can be viewed as a bit in an ECC, and redundant filters can be added to form parity check bits. This is also illustrated in Figure. 2, where three redundant filters are used to form the parity check bits of a classical single error correction Hamming code [14]. Those correspond to the outputs z1 [n], z2 [n], and z3 [n]. Errors can be detected by checking if

(2) When some of those checks fail, an error is detected. The error can be corrected based on which specifi checks failed. For example, an error on filter y 1 will cause errors on the checks of z1, z2, and z3

Figure 2. ECC-based scheme for four filters and a Hamming code.

CODING FOR FAULT-TOLERANT PARALLEL FILTERS The proposed scheme is illustrated in Figure. 3 for the case of four parallel filters. The input signals are encoded using a matrix with arbitrary coefficients to make the signals that enter the four original and two redundant filters. In its more general form, this coding matrix A can be formulated as

(3)

IJAAS Vol. 7, No. 4, December 2018: 353 – 360

IJAAS

ISSN: 2252-8814



357

Figure 3. Proposed coding scheme in general form

With this coding scheme, the outputs of the filters, i.e., y1[n], y2[n], y3[n], and y4[n], can be obtained as follows:

(4) where A1235 is a sub matrix of A, including the first, second, third, and fifth rows. This process can be repeated with different submatrixes of A, for example, with A1236, A2345, and A2346. In the error-free case, all the recovered versions of y1 [n], y2 [n], y3 [n], and y4 [n] will be the same. When there are differences, an error is detected. For xample, suppose that

(5) Which means that there is an error among filters {1 2 3 5 6} and that filters {2 3 4 5 6} are correct. Therefore, the faulty filter is filter 1. Then, the error can be corrected by taking the final outputs from a set that does not include filter 1. The error correction and detection logic can be simplified assuming that there is only a single error. In that case, checking only that, for each recovered set, the sums of the values y1 [n] + y2 [n] + y3 [n] + y4 [n] are equal is enough. In more detail, four checks are needed, each involving five filters and excluding one. For example, if branch 1 is excluded, the error checking would be

(6) in which w¯2345 = [1111](A2345) −1, and w¯2346 = [1111] (A2346) −1.

Coding Schemes for Implementation of Fault Tolerant Parrallel Filter (Nutenki Siddhartha)

358



ISSN: 2252-8814

RESULTS The written Verilog HDL Modules have successfully simulated and verified using Modelsim III 6.4b and synthesized using Xilinx ISE 13.2. Simulation Result

Figure 4. Simulation Result

In Figure 5 we observed that simulation results and Figure.6 and Figure.7 shows that RTL schematic and Technology schematic respectively.

Table 1. Design Summary Device Utilization Summary (Estimated Values) Logic Utilization Used Available Utilization Number of Slices 225 4656 4% Number of Slice FlipFLOPS Number of 4 input LUTs Number of Bonded IOBs Number of GCLKs

229

9312

421

9312

234

232

100%

RTL Schematic

Figure 5. RTL Schematic IJAAS Vol. 7, No. 4, December 2018: 353 – 360

IJAAS Technology Schematic

ISSN: 2252-8814



359

Figure 6. Technology Schematic

Timing Report

Figure 7. Timing Report 5.

CONCLUSION In this paper a new scheme to protect parallel filters that are commonly found in modern signal processing circuits has presented. The approach is based on applying ECCs to the parallel filters outputs to detect and correct errors. The technique can be used for parallel filters that have the same response and process different input signals. An objectives has also been discussed to show the effectiveness of the scheme in terms of error correction and problem definition also shows the overheads. The proposed scheme can also make system price lower. Proposed work will result in more efficient fault tolerant system using parallel IIR filters based on ECCs, which will meet the goal to achieve low power consumption, increase area of application and high speed.

REFERENCES [1] [2] [3] [4] [5] [6] [7]

P. P. Vaidyanathan, Multirate Systems and Filter Banks, Englewood Cliffs, N.J., USA: Prentice Hall, 1993. A. Sibille, C. Oestges and A. Zanella, MIMO: From Theory to Implementation, New York, NY, USA: Academic, 2010. N. Kanekawa, E. H. Ibe, T. Suga and Y. Uematsu, Dependability in Electronic Systems: Mitigation of Hardware Failures, Soft Errors, and ElectroMagnetic Disturbances, New York, NY, USA: Springer Verlag, 2010. M. Nicolaidis, “Design for soft error mitigation,” IEEE Trans. Device Mater. Rel., vol. 5, no. 3, pp. 405–418, Sep. 2005. C. L. Chen and M. Y. Hsiao, “Error-correcting codes for semiconductor memory applications: A state-of-the-art review,” IBM J. Res. Develop.vol. 28, no. 2, pp. 124–134, Mar. 1984. A. Reddy and P. Banarjee “Algorithm-based fault detection for signal processing applications,” IEEE Trans. Comput., vol. 39, no. 10, pp. 1304–1308, Oct. 1990. T. Hitana and A. K. Deb, “Bridging concurrent and non-concurrent error detection in FIR filters,” in Proc. Norchip Conf., 2004, pp. 75–78.

Coding Schemes for Implementation of Fault Tolerant Parrallel Filter (Nutenki Siddhartha)

360 [8] [9]



ISSN: 2252-8814

Y.-H. Huang, “High-efficiency soft-error-tolerant digital signal processing using fine -grain subword-detection Processing,” IEEE Trans. Very Large Scale Integr. (VLSI) Syst., vol. 18, no. 2, pp. 291–304, Feb. 2010. S. Pontarelli, G. C. Cardarilli, M. Re, and A. Salsano, “Totally fault tolerant RNS based FIR filters,” in Proc. IEEE IOLTS, Jul. 2008, pp. 192–194.

BIOGRAPHIES OF AUTHORS Nutenki Siddhartha is pursuing Master of Technology in Electronic Design Technology in SR Engineering College, Warangal, and Telangana. He has completed his B.Tech from Kamala Institute of Technology and Science, Karimnagar, Telangana in 2014. His areas of interests are VLSI, Digital Signal Processing and Computer Networks. He Published 1 Poster in ICRTEECT-2017.

G. Renuka working as assistant professor of Electronics and Communication Engineering, S R Engineering College, Warangal. She has 12 years of teaching experience. She has obtained B.Tech (ECE) Degree from Ramappa Engineering College Warangal, Andhra Pradesh, India in 2003 and M.Tech (Digital Communications) Degree from Kakatiya University, Andhra Pradesh, India in 2009. Pursuing her Ph.D under JNTUH. She is a member of IETE, and her research areas include VLSI, Communications.

IJAAS Vol. 7, No. 4, December 2018: 353 – 360

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 361~368 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp361-368



361

Lossless 4D Medical Images Compression Using Adaptive Inter Slices Filtering Leila Belhadef, Zoulikha Mekkakia Maaza Department d’Informatique, Laboratoire SIMPA, Faculté des Mathématiques et d’Informatique, Université des Sciences et de la Technologie d’Oran Mohamed Boudiaf, USTO-MB, BP 1505, El Mnaouer, 31000 Oran, Algérie

Article Info

ABSTRACT

Article history:

Recent lossless 4D medical images compression works are based on the application of techniques originated from video compression to efficiently eliminate redundancies in different dimensions of image. In this context we present a new approach of lossless 4D medical images compression which consists to application of 2D wavelet transform in spatial directions followed or not by either lifting transform or motion compensation in inter slices direction, the obtained slices are coded by 3D SPIHT. Our approach was compared with 3D SPIHT with/without motion compensation. The results show our approach offers better performance in lossless compression rate.

Received May 21, 2018 Revised Aug 5, 2018 Accepted Aug 29, 2018 Keyword: 3D SPIHT 4D medical image Integer wavelet transform Lossless compression Motion compensation

Corresponding Author: Leila Belhadef, Department d’Informatique, Laboratoire SIMPA, Faculté des Mathématiques et d’Informatique, Université des Sciences et de la Technologie d’Oran Mohamed Boudiaf, USTO-MB BP 1505, El Mnaouer, 31000 Oran, Algérie Email: leila.belhadef@univ-usto.dz

INTRODUCTION 4D medical images represent voluminous data where each image is composed by a set of volumes representing the 3D images of a human body part at different instants t, each 3D image is in turn composed by a set of slices. In the literature a lot of lossless 4D medical images compression works are realized to reduce images size without data loss in order to avoid diagnosis errors. Considering 4D image as a set of volumes evolving in time, the recent lossless 4D medical images compression techniques apply the techniques of video compression such as motion compensation in order to efficiently removing inter slices or volumes redundancies. Based on the advanced video coding H.264/AVC Sanchez et al. [1] described two coding methods for lossless 4D medical images compression. The first method exploits the redundancies among the 2D slices in (each volume) third dimension𝑧. The second method exploits the similarities between the slices in the fourth dimension time 𝑡. Redundancies through the fourth dimension are more numerous than redundancies through the third dimension; therefore this is the second method which permitted better compression ratio. For more exploit the redundancies in the four dimensions of medical images the same authors of the precedent work proposed a new compression technique based on H.264/AVC [2] as in precedent work. So in this compression technique multi-frame motion compensation is applied firstly in 𝑧 dimension to each volume. The obtained reference and residual slices are treated secondly also by multi-frame motion compensation in temporal dimension, thus the redundancies are reduced in 𝑧 and 𝑡 dimensions. The final residual slices and motion vectors are compressed by entropy coding. Here multi-frame motion compensation applied to a set of 𝑛 slices consists to consider first slice as reference slice intra coded and 𝑛 − 1 slices as inter slices which are inter coded by variable block matching Journal homepage: http://iaescore.com/journals/index.php/IJAAS

362



ISSN: 2252-8814

and multi-frame motion compensation where each block of inter slices can be predicted from multi reference slices in 𝑧 dimension in its first application and 𝑡 dimension in its second application. Martin et al. proposed in [3] a new method of 4D medical images compression also based on the H.264 video coding where predicted slices are generated by using spatio-temporal reference slices from 𝑧 and temporal neighborhood, where differently to precedent technique which realizes prediction in 𝑧 dimension followed by prediction in 𝑡 dimension, this technique realizes spatio-temporal prediction where each block of inter slice can be predicted from multi reference slices in spatio-temporal neighborhood, this permits to obtain a good results in compression compared prediction in 𝑧 dimension or prediction in time dimension. All the previous cited works (based on H.264 video coding) don’t integrate progressive lossy to lossless decoding in their schemes. Other techniques apply wavelet transform indeed it can be combined with image compression technique as fractal to improve time coding [4] also wavelet transform is very effective in medical image and video compression thus a lot of works are proposed as in [5] the authors presented color video medical compression technique by using geometric wavelet in order to eliminate efficiently similarity in each frame composed the video sequence with good image quality but without exploiting the temporal redundancies. The first proposed works based on wavelet transform in literature for lossless 4D medical images compression do not integrate motion compensation in compression scheme thus in [6] the authors compared three techniques using JPEG2000. The first technique consists to apply 2D wavelet transform to each slice composed the volumes and compress it separately by using JPEG2000. The second technique is represented by two variants: the first variant is application of 1D wavelet transform in 𝑧 dimension, the second variant is application of 1D wavelet transform in 𝑡 dimension and for both variants the obtained slices are compressed by JPEG2000. The third technique is application of 1D wavelet transform in 𝑡 dimension flowed by 1D wavelet transform in 𝑧 dimension the obtained slices are compressed by JPEG2000. The exploitation of temporal dimension in second and third technique provides better results in compression rate because the numerous similarities in temporal dimension. Due to efficiency of the motion compensation in the video compression scheme to reduce temporal redundancies Kasim et al. [7] proposed a wavelets compression technique of 4D medical images based on motion compensation and integer wavelet transform, here the 4D image is compressed as a sequence of 3D images. This technique consists to achieve 3D motion compensation in order to eliminate efficiently similarities existing between volumes in temporal direction, after the resulting volumes are decorrelated by 3D integer wavelet transform. The obtained data are coded by 3D SPIHT. This method allows lossless coding and progressive lossy to lossless decoding. The first step is motion estimation inter 3D images. Where a group of 𝑛 3D images are represented by one key image (the first image in a group is compressed without motion compensation) and 𝑛 − 1 inter images. Each current image (at position 𝑡) to be predicted is divided in cubes for each cube it most similar cube in reference image (at position 𝑡 − 1) is found. The difference between the position of current cube and it most similar cube is motion vectors and the difference between those cubes generates the residual image. Thus the results of this step are 3D key image, 𝑛 − 1 3D residual images and corresponding motion vectors. To reconstruct the initial group of images, the key image and the corresponding motion vectors is used to obtain predicted image which permits to reconstruct the first inter image by adding the predicted image to the first residual image, the followed inter images at position 𝑡 are reconstructed as the first one by using inter image at position 𝑡 − 1 and corresponding motion vectors. The second step consists to apply 3D integer wavelet transform to key image and residual images and in final step the key and residual images are coded by using 3D SPIHT, also the motion vectors are entropy coded. This technique permits to realize lossless compression and progressive lossy to lossless decoding by transmission a part of bit stream of key images followed by a part of bit stream of each inter image this permits to improve progressively the image quality in decompression. All cited works of 4D medical image compression try to eliminate the redundancies by using different prediction techniques in 𝑡 and 𝑧 dimensions either by applying motion compensation or by applying wavelet transform or by applying both motion compensation flowed by wavelet transform. In this paper we propose a new lossless compression approach based on adapted filtering in inter slices direction to improve the elimination of redundancies in 4D medical images. This approach consists to apply 2D integer wavelet transform to each slices followed or not by either wavelet filter or motion compensation in inter slices direction. The obtained slices are coded with 3D SPIHT. The experimental results are compared to two approaches: 3D SPIHT using motion compensated temporal filter in inter slices direction and 3D SPIHT (without motion compensation), the obtained lossless compression rates show our approach outperforms two other approaches. The rest of the paper is organized as follows. The section 2 presents prediction techniques used in video compression. The section 3 details the proposed approach of compression. In the section 4 the

IJAAS Vol. 7, No. 4, December 2018: 361 – 368

IJAAS

ISSN: 2252-8938



363

experimental results for lossless compression are compared to other compression approaches. In final section conclusions are presented.

MOTION COMPENSATION AND TEMPORAL FILTERING The principals video compression techniques used in 4D medical images compression are represented by two types of technique: H.264/AVC video coding and wavelet video coding. These techniques are based on motion compensation to eliminate the redundancies as in H.264/AVC coder the motion compensation is applied by using the block matching to attain the minimum prediction error which is the minimum difference between current block (to be predicted) and reference block in search window, also the reference blocks can be determined from multi reference images with variable block size. The lossless 4D compression techniques based on H.264/AVC offer good compression rate but without permitting the progressive decoding, the alternative is the wavelet techniques which produce scalable flux and realize good decorrelation of signal as presented in [7] where the progressive decoding is obtained by reordering of bit stream, the progressive decoding can be also obtained by using temporal scalability but it is not used here cause the motion compensation and wavelet transform is achieved separately. In order to obtain temporal scalability and efficiency eliminate temporal redundancies the recent wavelet video coding apply the wavelet transform in motion trajectory by using motion compensated temporal filtering (MCTF). In wavelet video coding the MCTF can be applied before spatial decorrelation with wavelet transform (t+2D) or after spatial decorrelation in wavelet domain (2D+t or in-band) flowed by entropy coding. The first works in video compression [8], [9] integrated Haar motion compensated temporal filtering where the images are separated in even images and odd images, the even images are low pass filtered and the odd images are high pass filtered, the obtained low pass images are also separated in even images and odd images and filtered till the temporal level decomposition is done. The filtering is realized in temporal trajectory determined by motion compensation to reduce the energy in obtained high pass filtered images. With the emergence of lifting scheme for the calculation of the integer wavelet coefficients, other works were carried as in [10] the authors defined two filters: motion compensated Haar lifting filter and motion compensated 5/3 lifting filter ; the use of motion compensated 5/3 lifting filter gave the best results. The scheme lifting is achieved using two steps prediction and update: the prediction step retains difference between two samples of signal (high pass filtering) and the update step retains the average therefore an approximation of signal (low pass filtering). In [11] the authors introduced the motion compensated truncated 5/3 lifting filter where only the prediction step of lifting is performed, thus each high pass filtered image is obtained by determining the prediction error from neighbor reference images forward and backward. We propose an adaptive prediction in order to ameliorate lossless 4D medical images compression rate, we apply either 5/3 truncated lifting filter or motion compensated, if one of them minimize prediction error in inter slices direction.

PROPOSED COMPRESSION SCHEME The first step of our scheme is formation of GOS (Group Of Slices) from 4D medical image followed by application of 2D integer wavelet transform (2D IWT) in spatial directions (𝑥, 𝑦) of each slices, the second step is inter slices filtering of each GOS and the final step is coding of obtained slices, motion vectors and etiquettes. As shown in Figure 1 proposed compression scheme.

Figure 1. Proposed compression scheme We describe these steps in the following points.

Lossless 4D Medical Images Compression Using Adaptive Inter Slices Filtering (Leila Belhadef)

364



ISSN: 2252-8814

3.1. Construction of GOS and spatial transform The GOS are formed from the slices of 4D medical image (𝑥,𝑦,𝑧,𝑡) as shows in Figure 2(a). This construction permits to obtain a set of GOS representing 4D medical image with each GOS is composed with nearest slices in spatial direction 𝑧 across time.

Figure 2. (a) 4D Image composed by 3 volumes. (b) Construction of GOS composed by 16 slices (in gray)

In Figure 2(b) the slices in gray represent the first GOS with 16 slices of 4D image composed by 3 volumes, thus all GOS of 4D image are also formed with follow slices in the same direction used for the first GOS (across time), as the second GOS starts with slice 17. Next the formation of GOS each slice is transformed by 2D integer wavelet transform using lifting scheme [12] in spatial directions (𝑥, 𝑦). This transform is reversible; it allows producing integer wavelet coefficients which permits lossless compression. 3.2. Inter Slices Filtering After spatial transform each obtained GOS is filtered by the proposed filter in inter slices direction based on truncated lifting scheme and motion compensation. The motion compensated truncated 5/3 lifting filter proposed in [11] is realized by application of the truncated lifting scheme and motion compensation in the same time, the truncated lifting scheme achieves only prediction step of lifting scheme. The motion compensated truncated 5/3 lifting filter (MC Trunc 5/3) for the block Sk [m, n] can be formulated as: 𝑅[𝑚, 𝑛] = 𝑆𝑘 [𝑚, 𝑛] − ⌊0.5 × (𝑆𝑘−1 [𝑚 − 𝑑1𝑚 , 𝑛 − 𝑑1𝑛 ] + 𝑆𝑘+1 [𝑚 − 𝑑2𝑚 , 𝑛 − 𝑑2𝑛 ])⌋

(1)

Where Sk-1, Sk+1 : reference slices, and Sk: inter slice. R: residual slice (high pass slice). (d1m, d1n): motion vector of block Sk [m, n] to a position in Sk-1. (d2m, d2n): motion vector of block Sk [m, n] to a position in Sk+1.

. : corresponds to round operator.

The motion compensation is realised by using block motion model so each inter slice (to be predicted) is divided in blocks and each block is filtered by MC Trunc 5/3. This filter achieves bi-directional motion compensation thus for each current block in Sk it most similar block is determined in Sk-1 and in Sk+1 and the difference between the position of current block and these blocks represents motion vectors, the criteria of similarities used is minimum sum of absolute differences (SAD). Thus each inter slice is replaced by two motion fields as shows in Figure 3(a) and residual slice.

IJAAS Vol. 7, No. 4, December 2018: 361 – 368

IJAAS



ISSN: 2252-8938

(a)

365

(b)

Figure 3. Motion fields: (a) predicted block with MC Trunc 5/3, (b) proposed approach with 1: unfiltered block(E1), 2: predicted block with truncated 5/3 lifting transform(E2), 3: predicted block with motion compensated(E3)

In our scheme each inter slice (to be predicted) is divided in block and each block is either unfiltered or filtered by one of following prediction techniques: Truncated 5/3 lifting transform or motion compensation. This filtering can be represented by three following prediction errors: The unfiltered block Sk[m,n]: 𝐸1 [𝑚, 𝑛] = 𝑆𝑘 [𝑚, 𝑛]

(2)

The truncated 5/3 lifting transform (without motion compensation): 𝐸2 [𝑚, 𝑛] = 𝑆𝑘 [𝑚, 𝑛] − ⌊0.5 × (𝑆𝑘−1 [𝑚, 𝑛] + 𝑆𝑘+1 [𝑚, 𝑛])⌋

(3)

The motion compensated prediction: 𝐸3 [𝑚, 𝑛] = 𝑆𝑘 [𝑚, 𝑛] − 𝑆𝑘−1 [𝑚 − 𝑑𝑚 , 𝑛 − 𝑑𝑛 ]

(4)

Where (dm, dn): motion vector of block Sk [m, n] to a position in Sk-1. Between the three prediction errors the minimum is chosen for determining the prediction error of this block Sk[m,n] and its prediction method. The prediction error E1 represents the unfiltered block Sk[m,n], the prediction error E2 represents prediction step in 5/3 lifting transform and the prediction error E3 is the difference between block Sk[m,n] and it most similar block in slice Sk-1 as shows in Figure 3(b), the motion vector (dm, dn) represents the displacement which allows minimizing E3. So each inter slice is replaced by one motion field (forward) and/or etiquette (to indicate which prediction error is used) and residual slice.

Figure 4. 3 levels motion compensated truncated 5/3 lifting filter for GOS with 16 slices Lossless 4D Medical Images Compression Using Adaptive Inter Slices Filtering (Leila Belhadef)

366



ISSN: 2252-8814

Figure 4 presents 3 decomposition levels of MC Trunc 5/3 for GOS with 16 slices; the number of obtained slices of MC Trunc 5/3 is the same to our scheme which is composed by 2 reference slices (1, 9) and 14 residual slices (R1, R2, … R14), the slice (17) represents the first reference slice in the next GOS. However for motion coding, the approach with MC Trunc 5/3 produces two motion fields and our approach one motion field (forward) and etiquette for each predicted block by E3 and only etiquette for each predicted block by E1 or E2. 3.3. Reference and residual slices, motion vectors and etiquettes coding The obtained reference and residual slices composed GOS are coded by 3D SPIHT [13] (Set Partitioning in Hierarchical Trees) and generated bit stream is coded by arithmetic coder, 3D SPIHT is used as in video coding and volumetric image coding, it is a bit plan coder and it permits to reduce data size by exploiting the inter dependences of subbands in three dimensions 𝑥, 𝑦 and inter slices. Concerning the motion vectors and etiquettes are coded by arithmetic coder.

EXPERIMENTAL RESULTS We tested the proposed compression method with five native 4D medical images CT (Computed Tomography) of heart from two references [14][15] shown in Figure 5. Each one is composed by 10 volumes and the spatial resolution of slices is 512×512 coded on 16bits per pixel. However the volumes of each 4D medical image have different sizes: data of reference [14] Image1 (141 slices/volume), Image2 (169 slices/volume) and Image3 (170 slices/volume) data of reference [15] Image4 (136 slices/volume) and Image5 (120 slices/volume).

Figure 5. First slices of each 4D medical image

We compare the experimental results of our approach with MC Trunc 5/3 approach (3D SPIHT with MC Trunc 5/3) and 3D SPIHT without motion compensation. For all these approaches GOS are composed by 16 slices formed as shows in Figure 2 (b). IJAAS Vol. 7, No. 4, December 2018: 361 – 368

IJAAS



ISSN: 2252-8938

367

After each GOS is transformed with 3 levels 2D integer wavelet transform in spatial directions (𝑥, 𝑦) we use 5/3 filter. In inter slices direction of GOS our approach uses the proposed filter equations (2, 3 and 4), the MC Trunc 5/3 approach uses equation (1) with also three levels of decomposition for both. However 3D SPIHT without motion compensation uses in inter slices direction 1D integer wavelet transform with three levels of decomposition of the same used filter in spatial directions (5/3). The obtained subbands of three approaches are coded with 3D SPIHT. The lifting scheme is very applied in medical image compression with traditionnel filters as 5/3 but other new filters can be tested in spatial directions [16]. The motion vectors are determined by block motion compensation model with full search method, the search is effectuated with block size of 16×16 pixels in our approach and MC Trunc 5/3 approach. The Table 1 lists the obtained average bit rate for the lossless compression using two first volumes of each 4D medical image.

Table 1. 4D Lossless Compression Results in bit per pixel (bpp) 4D Medical Images Image1 Image2 Image3 Image4 Image5

3D SPIHT 5.19 5.92 5.39 5.03 4.53

MC Trunc 5/3 5.23 5.97 5.45 5.06 4.55

Proposed Method 5.10 5.89 5.36 4.98 4.35

The results show our method outperforms the other methods for all test images. 3D SPIHT gives low average bit rate than MC Trunc 5/3 of 0.76% , thus the integration of the motion compensated truncated 5/3 lifting filter in 3D SPIHT penalises the compression rate in MC Trunc 5/3 approach. However the proposed inter slices filter improves the compression rate (low bit rate) with average of 1.45% compared to 3D SPIHT, as well a net improvement is reached of the order of 2.2% compared to MC Trunc 5/3. The average improvement by proposed technique compared to 3D SPIHT is obtained by using image1 thus we use it in following test. The experimental results presented in Table 2 are obtained by testing five different sizes of image1, thus line 2 represents lossless compression results of 2 first volumes of image1, line 3 of 4 first volumes of Image1 and so on to the last line of 10 first volumes of image1, these different sizes of Image1 permit to obtain differents GOS by reorganizing the slices as in Figure 2 (b).

Table 2. 4D Lossless Compression Results in bit per pixel (bpp) for different sizes of Image1 4D Medical Images 2Volumes Image1 4Volumes Image1 6Volumes Image1 8Volumes Image1 10Volumes Image1

3D SPIHT 5.19 5.08 5.07 4.98 4.99

MC Trunc 5/3 5.23 5.10 5.05 5.00 4.96

Proposed Method 5.10 4.99 4.96 4.92 4.90

The obtained compression rate by proposed method is better than other schemes for all images test. We note that the bit rate is reduced for all compression schemes while the number of temporal slices (at different time and with the same position 𝑧) composed GOS increase. The results show a variation of the compression rate as a function of the number of volumes. The improvement of this rate by our approach is of the order of 1.73% and 1.85% compared to 3D SPIHT and MC Trunc 5/3 respectively, while it does not exceed 0.11% between the latter two methods.

CONCLUSION In this paper we proposed a new approach of lossless compression of 4D medical images, it consists in construction of GOS and each slices composed GOS is decorrelated in spatial directions (𝑥,𝑦), after proposed inter slices filtering is performed with minimum of three predicted errors for each block composed inter slices. The three prediction errors are represented by unfiltered block, truncated 5/3 lifting filter and motion compensated the aim is to reduce the size of residual slices. The obtained slices are coded with 3D SPIHT. The proposed approach provides lossless compression improvements by reducing the bit rates compared to 3D SPIHT with MC trunc 5/3 and 3D SPIHT without motion compensation, this enhancement is due to the integration of the three prediction techniques in inter slices filter. As future work, our approach offers the possibility to obtain lossy to lossless decoding by exploiting inter slices scalability also other 4D medical image modalities can be tested. Lossless 4D Medical Images Compression Using Adaptive Inter Slices Filtering (Leila Belhadef)

368



ISSN: 2252-8814

REFERENCES [1] V. Sanchez, P. Nasiopoulos, R. Abugharbieh, "Lossless Compression of 4D Medical Images using H.264/AVC, " IEEE International Conference on Acoustics, Speech, and Signal Processing, Toulouse, 2006, pp. 1116-1119. [2] V. Sanchez, P. Nasiopoulos, R. Abugharbieh, "Efficient lossless compression of 4D medical images based on the advanced video coding scheme," IEEE Trans. Information Technology In Biomedicine, Vol. 12, NO. 4, JULY 2008, pp. 132-138. [3] U. Martin, A. Kaup, "Analysis of compression of 4D volumetric medical image datasets using multi-view (MVC) video coding methods," Mathematics of Data/Image Pattern Recognition, Compression and Encryption with Applications XI, Proc. of SPIE. Vol. 7075, 70757,2008. [4] Y. Feng, H. Lu, X. Zeng, "A Fractal Image Compression Method Based on Multi-Wavelet", TELKOMNIKA (Telecommunication, Computing, Electronics and Control), Vol.13, No.3, September 2015, pp. 996-1005. [5] Y. Habchi, M. Beladgham, A. A. Taleb , "RGB Medical Video Compression Using Geometric Wavelet and SPIHT Coding ", International Journal of Electrical and Computer Engineering (IJECE), Vol. 6, No. 4, August 2016, pp. 1627-1636. [6] H. G. Lalgudi, A. Bilgin, M. W. Marcellin, A. Tabesh, M. D. Nadar and T. P. Trouard, "Four-dimensional compression of fMRI using JPEG2000," Medical Imaging 2005:Image Processing, Proc. of SPIE. Vol. 5747. [7] A. A. Kassim, P. Yan, W. S. Lee, K. Sengupta, "Motion Compensated Lossy-to-Lossless Compression of 4D Medical Images Using Integer Wavelet Transforms," IEEE Trans. Information Technology In Biomedicine, VOL. 9, NO. 1, 2005, pp. 132-138. [8] J. -R. Ohm, "Three-dimensional subband coding with motion compensation," IEEE Trans. on Image Processing, 1994, pp. 559-571. [9] S. -J. Choi and J. Woods, "Motion-compensated 3-d subband coding of video," IEEE Trans. on Image Processing, 1999, pp. 155-167. [10] A. Secker and D. Taubman, "Motion-compensated highly scalable video compression using an adaptative 3d wavelet transform based on lifting," IEEE, 2001. [11] L. Luo, J. Li, S. Li, Z. Zhuang, and Y-Q .Zhang, "Motion-compensated lifting wavelet and its application in video coding," in IEEE International Conference on Multimedia and Expo, August 2001. [12] I. Daubechies and W. Sweldens, "Factoring wavelet transforms into lifting steps," Journal of Fourier Analysis and Applications, 1998. [13] B. J. Kim and W. A. Pearlman, "An Embedded Wavelet Video Coder Using Three-Dimensional Set Partitioning in Hierarchical Trees (SPIHT)," In IEEE Data Compression Conference, 1997, pp. 221-260. [14] J. Vandemeulebroucke, S. Rit, J. Kybic, P. Clarysse, and D. Sarrut, "Spatiotemporal motion estimation for respiratorycorrelated imaging of the lungs," In Med Phys, 2011, pp. 166-178. [15] http://midas.kitware.com/community/view/47. [16] A. Hazarathaiah, B. Prabhakara Rao, "Medical Image Compression using Lifting based New Wavelet Transforms", International Journal of Electrical and Computer Engineering (IJECE), Vol. 4, No. 5, October 2014, pp. 741-750.

IJAAS Vol. 7, No. 4, December 2018: 361 – 368

International Journal of Advances in Applied Sciences (IJAAS) Vol. 7, No. 4, December 2018, pp. 369~376 ISSN: 2252-8814, DOI: 10.11591/ijaas.v7.i4.pp369-376



369

A Secure Data Transmission Scheme using Asymmetric SemiHomomorphic Encryption Scheme S. Nagavalli1, G.Ramachandran2 of Computer and Information Sciences, Annamalai University, Annamalainagar –608 002, Tamil Nadu, India. 2Department of Computer Science & Engineering, Annamalai University, Annamalainagar –608 002, Tamil Nadu, India. 1Department

Article Info

ABSTRACT

Article history:

The compressive detecting based information accumulation accomplishes with high exactness in information recuperation from less inspection which is available in sensor nodes. In this manner, the existing methods available in the literature diminish the information gathering cost and delays the existence cycle of WSNs. In this paper, a strong achievable security model for sensor network applications was initially proposed. At that point, a secure data collection conspire was displayed based on compressive detecting, which improves the information protection by the asymmetric semi-homomorphic encryption scheme, and decreases the calculation cost by inadequate compressive grid. In this case, particularly the asymmetric mechanism decreases the trouble of mystery key circulation and administration. The proposed homomorphic encryption permits the in-arrange accumulation in cipher domain, and in this manner improves the security and accomplishes the adjustment in system stack. Further, this paper focuses on estimating various network performances such as the calculation cost and correspondence cost, which remunerates the expanding cost caused by the homomorphic encryption. A real time validation on the proposed encryption scheme using AVISPA was additionally performed and the results are satisfactory.

Received May 28, 2018 Revised Jul 26, 2018 Accepted Aug 21, 2018 Keyword: Homomorphic Encryption Symmetric Cryptosystem Wireless Sensor Network

Corresponding Author: S. Nagavalli, Department of Computer and Information Sciences, Annamalai University, Annamalainagar–608 002, Tamil Nadu, India. Email: nagavalli52@gmail.com

INTRODUCTION Wireless Sensor Networks (WSN) has been deployed in different applications which include personal, business and military regions. It comprises of bunches of minimal effort and battery controlled hubs, which are regularly circulated in unattended conditions. Secure information accumulation has been proposed as a pivotal way to deal with settling the vitality and security challenges. The compressive detecting based information accumulation [1-3], which join information procurement with information pressure, can outperform the limits of the customary hypothesis by investigating the sparsity of compressible signs. It accomplishes high accuracy information recuperation from less testing information, and in this way diminishes the information accumulation cost and draws out the existence cycle of WSNs. In recent times, most of the sensor hubs can't avoid assaults and are anything but difficult to catch, so hubs are not appropriate for put away private data. Even attackers with their assault models are efficient; the enemy traded off a hub initially, and afterward actualized the assault to get the estimation grid. In order to secure the sensor network, various minimal encryption schemes have been proposed. Numerous works on sensor crypto system are being carried out [5, 6] based on the randomized estimation of security conservation. The Symmetric Cryptosystem [4] was first clarified as a single key cryptosystem, where the pseudo-random Journal homepage: http://iaescore.com/journals/index.php/IJAAS

370



ISSN: 2252-8814

estimation framework was used as the key, and the estimation result is the figure content of unique signs. It is a light-weight encryption plot. The encoding step and the recouping step can be translated as the encryption and the unscrambling separately, and along these lines no additional computational cost is required. The controllable occasion activating assault situation is just changed the estimation of one hub, which means the irregular occasion activating assault model has been changed and all the assaults are targeted to specific nodes. Hence, the existing encryption system which uses the typical private (mystery) key cannot be reused. In order to address the above mentioned issues, this paper proposes a novel homomorphic encryption scheme. The key idea is to guarantee the refresh of key and proof based hash capacities, which assumes the nodes are secure by themselves. The contribution of this paper is as follows: 1. Proposed a minimalistic encryption standard which supports all the sensor network model. 2. Compressive detecting based encryption scheme with support to optimal securing strategy. 3. A secure proof based on hash capacities which assumes the nodes are secure. The rest of this paper is organized as follows: Section II discusses the state of the art methods on homomorphic encryption schemes. Section III deals with the proposed encryption scheme. Section IV discusses about the experimentation of encryption schemes. The penultimate section of this paper discusses about the performance analysis and finally the paper is concluded.

LITERATURE SURVEY In this section, various state of the art techniques available in the literature are discussed. This section dissects into three phases namely i) Obscurity ii) Directing consolidated plan iii) Destructive modification. 2.1. Obscurity In the obscurity based plan, a straight change is connected to the first information, where the change grid is the arbitrary grid. This straight change is deciphered as a lightweight symmetric encryption plot, where the arbitrary grid is the private key. In the compressive detecting encryption plans [7, 8, 11], the vector size of the change comes about is substantially less than that of the first information, which implies the first information is packed amid the change. In the randomize change based plan [9, 10], the confusion change is connected to the compressive estimation comes about, and the vector size of the estimation comes about is the same as that of the change result. 2.2. Directing consolidated plan In these plans, each source hub partitions the first information into a few sections and transmits them in multi-way, while the middle of the road hubs makes some further procedure. In cut blend based plans [12], each source hub parts its unique information into a few cuts, and encodes them before sending them out in multipath. The moderate hub will unscramble these cuts and total them together. In secure system coding based plan [13], each source hub isolates the first information into a few squares what's more, dirties some of them. These squares are conveyed in multipath, what's more, will be re-encoded in the moderate hub. Each source hub additionally produces a few kinds of CRCs and encodes them. The security of square information is accomplished through the contamination component, which is in reality a sort of confusion system. Without unscrambling these CRCs, the foe is as yet conceivable to recuperation unique squares in the event that got enough bundles. 2.3. Destructive modification based scheme In this kind of plan, the information protection is accomplished by applying a ruinous alteration on the first information. For instance, the information security of the differential protection based plan [7] is safeguarded by differential security hypothesis, where the arbitrary clamor is added to decrease the information certainty. The amusement hypothesis based plans [5,8] contemplate the entrance of private information by utilizing monetary approaches, for example, amusement hypothesis and contract hypothesis. Each source decides its security level and changes its information in view of the harmony thought of security and reward. Information gathered by these systems is regularly utilized for promote information mining. 2.4. Routing Schemes Prasad et.al [16] presented different routing techniques in the IoT. An attack model is used for routing because of security issues in the IoT. Some of the attacks re-route the packets to the attacker’s controller. Through this information, the routing communication is extracted. Most common attacks are DDoS, Traffic hijacking, which exploits routing mechanism. Paul et.al [17] presented a multi hop protocol which uses multiple parameters for secure routing which have particular information already known as predefined for IJAAS Vol. 7, No. 4, December 2018: 369 – 376

IJAAS

ISSN: 2252-8938



371

users. Resources are stored as routing information. Results are showed that secure multi-hop routing mechanism had been used for IoT communication. Saleem et.al [15] proposed a bio-inspired secure IPv6 communication protocol for IoT. They enhanced the lossy network and low power transmission by classification algorithm called artificial immune system which classifies the misbehaving nodes and normal nodes with local information. Through the classification detection of excessive broadcast, improvements in power and transmission rate are increased. Results proved that transmission rate and energy consumption is far better than previously state of art routing techniques of IoT. Liu et.al [14] proposed a SDN based IoT secure routing protocol. The issue in these IoT routing mechanism with SDN is security middlebox guard. SDN security based data transfer security model is reducing network latency and manage to secure data flow with the help of heuristic algorithms. Middle boxes with secure policies are placed at different locations. Next to tackle against changing of middleboxes to honeypot. They use offline integer program and also it is used to load balance. Experiments are demonstrated that this model can handle secure routing mechanism for IoT.

PROPOSED ENCRYPTION SCHEME Compressive detecting based encryption scheme is basically an asymmetric encryption scheme. Figure 1 shows the architecture of the proposed encryption scheme. As per the matter of the fact, asymmetric encryption schemes are costlier for WSN applications when compared to other encryption schemes. However, these schemes can be made into minimalistic and can be deployed in sensor applications for better security. This can be achieved by limiting the security parameter sizes. Choosing the number of security parameters can also make the schemes feasible for deployment. This happens at the expense of the security of the scheme, but it was found that it still provides an appropriate level of security. In the proposed encryption scheme, the public parameters chosen are a larger positive integer g which is 10200 and a positive integer d which is greater than two, as per the assumption. Hence the integer g should always have many small divisors and also many integers which are small when it is inverted modulo g. Consider, the following case for validation. 1. The first case proposed is that d should not be greater than 4 and should include the lower bound 2. 2. The second case is that g should not be greater than 232. The list of parameters are always a positive integer r ∈ Zg and a positive integer g` which include logg, g is a always a secret security parameter. Finally, the secret key of the scheme is defined as (r, g`). Hence to encrypt the data from the WSN, m ∈ Zg`, where d is always a random number (S1 to Sd) should be generated 𝑚 = ∑𝑑𝑗=1 𝑠𝑗 𝑚𝑜𝑑 𝑔′ and sj > Zg. Further, the cipher text is then found as 𝐸(𝑚) = ∑𝑑𝑗=1 𝑠𝑗 𝑚𝑜𝑑 𝑔′. Additionally, addition and subtraction are to be done component wise where multiplication and division is done by multiplying the components of Zg.

A Secure Data Transmission Scheme using Asymmetric Semi-Homomorphic… (S. Nagavalli)

372



ISSN: 2252-8814 1

AES - 128

Network controller

create uses

RSA - 512

Device public key

inclusion message

CRC check sum

encrypted with

RSA - 512

Controller key AES - 128 to encrypt

AES - 128

3. Transmitted to device

Complete encryptred message

Network key

4. computed on the device

decrypted with

RSA - 512

AES - 128 Device private key

Decrypted network key

Figure 1. Architecture of the Proposed Crypto-System for WSN

EXPERIMENTAL SETUP The proposed encryption scheme is implemented on Node Red and the nodes are deployed independently to monitor the water tank which contains shrimps. Each node is executed to run encryption scheme in order to ensure that each node communication is encrypted. A checksum is created and certificate pinning is also ensured in all nodes in order to verify the authenticity of nodes and to avoid the node cloning activities. The entire scheme is implemented using python. Figure 2 shows the architecture of sensor. Table 1 shows the data aggregation rate for various size of network.

Onboard memory

Main processing unit Analog to Digital converter

Onboard cryptoprocessing system

External memory

Transceiver

Figure 2. Architecture of the Sensor Deployment IJAAS Vol. 7, No. 4, December 2018: 369 – 376

Sensing environment

IJAAS



ISSN: 2252-8938

373

Table 1. Encrypted Traffic Analyzed for Various Iteration with Increased Number of Nodes Node 1 2 3 4

Plain text Trans Receiv 50 20 52 30 54 34 60 45

Node 1 2 3 4

Trans 70 72 74 80

Plain text Receiv 30 40 54 65

Node 1 2 3 4

Trans 80 97 99 110

Plain text Receiv 40 55 67 76

25 nodes Round 1 Trans Receiv 110 70 120 80 134 90 154 110 50 nodes Round 1 Trans Receiv 140 80 150 99 164 119 184 140 75 nodes Round 1 Trans Receiv 156 99 167 112 187 123 198 154

Trans 120 134 140 160

Round 2 Receiv 75 82 94 125

Trans 131 151 166 172

Round 3 Receiv 70 98 110 120

Trans 150 164 170 180

Round 2 Receiv 87 110 123 190

Trans 151 167 178 197

Round 3 Receiv 80 112 132 145

Trans 166 176 187 199

Round 2 Receiv 90 112 132 212

Trans 165 176 187 210

Round 3 Receiv 100 123 154 198

4.1. Performance Analysis The following parameters are selected for evaluation of the proposed asymmetric encryption algorithms for both encryption and decryption schemes. 1. Encryption Time The encryption time is considered as the time that an encryption algorithm takes to produces a ciphertext from a plain text. Fig. 4 and Fig. 6 shows the encryption time of various file size. 2. Decryption Time The decryption time is considered as the time that a decryption algorithm takes to reproduces a plain text from a ciphertext. Figure 3 and Figure 5 shows the decryption time of various file size. 3. Throughput Throughput is equal to total plaintext in bytes encrypted divided by the encryption time. Higher the throughput, higher will be the performance. 4. Encrypted File Size The size of encrypted file is called encrypted file size. 5. Decrypted File Size The size of decrypted file is called decrypted file size.

Figure 3. Decryption File Size (Proposed Homomorphic Scheme vs Existing Homomorphic Scheme) A Secure Data Transmission Scheme using Asymmetric Semi-Homomorphic… (S. Nagavalli)

374



ISSN: 2252-8814

Figure 4. Encryption File Size (Proposed Homomorphic Scheme vs Existing Homomorphic Scheme)

Figure 5. Decryption Time (Proposed Homomorphic Scheme vs Existing Homomorphic Scheme)

IJAAS Vol. 7, No. 4, December 2018: 369 – 376

IJAAS

ISSN: 2252-8938



375

Figure 6. Encryption Time (Proposed Homomorphic Scheme vs Existing Homomorphic Scheme)

CONCLUSION WSN is popular in increasingly gaining popularity in most of the applications in various domains. In order to provide security to the sensor node communication, this paper presented a homomorphic encryption scheme which allows data processing on a node encryption. Further it is also found that the proposed encryption scheme is for all sorts of networks. Even if the network size grows, the proposed scheme is capable to work on. It was also found that the proposed encryption scheme does not significantly reduce the performance of plaintext aggregation. This means that the proposed encryption scheme is feasible for WSN applications while considering its effect on network traffic.

REFERENCES J. H. Cheon et al., “Toward a Secure Drone System: Flying with Real-Time Homomorphic Authenticated Encryption”, In IEEE Access, vol. 6, pp. 24325-24339, 2018. [2] H. Chen, Y. Hu and Z. Lian, “Leveled Homomorphic Encryption in Certificateless Cryptosystem”, In Chinese Journal of Electronics, vol. 26, no. 6, pp. 1213-1220, 2017. [3] H. Chen, Y. Hu and Z. Lian, “Properties of SV-Style Homomorphic Encryption and Their Application”, In Chinese Journal of Electronics, vol. 26, no. 5, pp. 926-932, 2017. [4] B. Wang, Y. Zhan and Z. Zhang, “Cryptanalysis of a Symmetric Fully Homomorphic Encryption Scheme”, In IEEE Transactions on Information Forensics and Security, vol. 13, no. 6, pp. 1460-1467, June 2018. [5] L. T. Phong, Y. Aono, T. Hayashi, L. Wang and S. Moriai, “Privacy-Preserving Deep Learning via Additively Homomorphic Encryption”, In IEEE Transactions on Information Forensics and Security, vol. 13, no. 5, pp. 13331345, May 2018. [6] Alabdulatif, H. Kumarage, I. Khalil, M. Atiquzzaman and X. Yi, “Privacy-preserving cloud-based billing with lightweight homomorphic encryption for sensor-enabled smart grid infrastructure”, In IET Wireless Sensor Systems, vol. 7, no. 6, pp. 182-190, 12 2017. [7] L. Chen, M. Lim and Z. Fan, “A Public Key Compression Scheme for Fully Homomorphic Encryption Based on Quadratic Parameters With Correction”, In IEEE Access, vol. 5, pp. 17692-17700, 2017. [8] K. Lauter, “Postquantum Opportunities: Lattices, Homomorphic Encryption, and Supersingular Isogeny Graphs”, In IEEE Security & Privacy, vol. 15, no. 4, pp. 22-27, 2017. [9] Y. Ma, L. Wu, X. Gu, J. He and Z. Yang, “A Secure Face-Verification Scheme Based on Homomorphic Encryption and Deep Neural Networks”, In IEEE Access, vol. 5, pp. 16532-16538, 2017. [10] N. Dowlin, R. Gilad-Bachrach, K. Laine, K. Lauter, M. Naehrig and J. Wernsing, “Manual for Using Homomorphic Encryption for Bioinformatics”, In Proceedings of the IEEE, vol. 105, no. 3, pp. 552-567, March 2017. [1]

A Secure Data Transmission Scheme using Asymmetric Semi-Homomorphic… (S. Nagavalli)

376



ISSN: 2252-8814

[11] Khedr and G. Gulak, “SecureMed: Secure Medical Computation Using GPU-Accelerated Homomorphic Encryption Scheme”, In IEEE Journal of Biomedical and Health Informatics, vol. 22, no. 2, pp. 597-606, March 2018. [12] M. S. Lee, “Sparse subset sum problem from Gentry–Halevi's fully homomorphic encryption”, In IET Information Security, vol. 11, no. 1, pp. 34-37, 2017. [13] R. Bocu and C. Costache, “A homomorphic encryption-based system for securely managing personal health metrics data”, In IBM Journal of Research and Development, vol. 62, no. 1, pp. 1:1-1:10, Jan.-Feb. 1 2018. [14] Y. Liu, Y. Kuang, Y. Xiao and G. Xu, “SDN-Based Data Transfer Security for Internet of Things”, In IEEE Internet of Things Journal, vol. 5, no. 1, pp. 257-268, Feb. 2018. [15] K. Saleem, J. Chaudhry, M. A. Orgun and J. Al-Muhtadi, “A bio-inspired secure IPv6 communication protocol for Internet of Things”, 2017 Eleventh International Conference on Sensing Technology (ICST), pp. 1-6, Sydney, NSW, 2017. [16] Prasad, Shyam Sundar, and Chanakya Kumar, “An energy efficient and reliable internet of things”, In Communication, Information & Computing Technology (ICCICT), pp. 1-4, IEEE, 2012. [17] Paul AS Ward, Evan PC, Martin Karsten, and Jones, “Multipath load balancing in multi-hop wireless networks”, In Wireless And Mobile Computing, Networking And Communications, (WiMob'2005), IEEE International Conference, vol. 2, pp. 158-166, 2005.

BIOGRAPHIES OF AUTHORS Dr.G.Ramachandran received the B.E degree in Computer Science and Engineering from Annamalai University in 1997. He received the M.E degree in Computer Science and Engineering from Annamalai University in the year 2005. He has been with Annamalai University, since 2000. He completed his Ph.D degree in Computer Science and Engineering at Annamalai University, in the year 2014. He published 30 papers in International conferences and Journals. His research interest includes Computer Networks, Network Security, Wireless Networks, Mobile Ad hoc networks and IoT.

S. Nagavalli received the B.Sc degree in Computer Science from CKN college in 2002. She completed her M.C.A degree from AVC college in the year 2005. She received her M.Phil degree from Annamalai University in 2008. She is doing her Ph.D degree in Department of Computer and Information Science at Annamalai University. She has published 3 papers in International Journals. Her area of interest includes Internet of Things, Cryptography, and Computer Networks.

IJAAS Vol. 7, No. 4, December 2018: 369 – 376

Institute of Advanced Engineering and Science Indonesia

: D2, Griya Ngoto Asri, Bangunharjo, Sewon, Yogyakarta 55187, Indonesia

Malaysia

: 51 Jalan TU 17, Taman Tasik Utama, 75450 Malacca, Malaysia

COPYRIGHT TRANSFER FORM (Please compile this form, sign and send by e-mail) Please complete and sign this form and send it back to us with the final version of your manuscript. It is required to obtain a written confirmation from authors in order to acquire copyrights for papers published in the International Journal of Advances in Applied

Sciences (IJAAS) Full Name and Title Organisation Address and postal code City Country Telephone/Fax E-mail Paper Title Authors

Copyright Transfer Statement The copyright to this article is transferred to Institute of Advanced Engineering and Science (IAES) if and when the article is accepted for publication. The undersigned hereby transfers any and all rights in and to the paper including without limitation all copyrights to IAES. The undersigned hereby represents and warrants that the paper is original and that he/she is the author of the paper, except for material that is clearly identified as to its original source, with permission notices from the copyright owners where required. The undersigned represents that he/she has the power and authority to make and execute this assignment. We declare that: 1. This paper has not been published in the same form elsewhere. 2. It will not be submitted anywhere else for publication prior to acceptance/rejection by this Journal. 3. A copyright permission is obtained for materials published elsewhere and which require this permission for reproduction. Furthermore, I/We hereby transfer the unlimited rights of publication of the above mentioned paper in whole to IAES. The copyright transfer covers the exclusive right to reproduce and distribute the article, including reprints, translations, photographic reproductions, microform, electronic form (offline, online) or any other reproductions of similar nature. The corresponding author signs for and accepts responsibility for releasing this material on behalf of any and all co-authors. This agreement is to be signed by at least one of the authors who have obtained the assent of the co-author(s) where applicable. After submission of this agreement signed by the corresponding author, changes of authorship or in the order of the authors listed will not be accepted.

Retained Rights/Terms and Conditions 1. 2.

Authors retain all proprietary rights in any process, procedure, or article of manufacture described in the Work. Authors may reproduce or authorize others to reproduce the Work or derivative works for the author’s personal use or for company use, provided that the source and the IAES copyright notice are indicated, the copies are not used in any way that implies IAES endorsement of a product or service of any employer, and the copies themselves are not offered for sale. Although authors are permitted to re-use all or portions of the Work in other works, this does not include granting third-party requests for reprinting, republishing, or other types of re-use.

Yours Sincerely,

Corresponding Author‘s Full Name and Signature Date: ……./……./…………

IAES Journals, http://iaescore.com/journals International Journal of Advances in Applied Sciences (IJAAS), email: iaes.editor@gmail.com, editor@iaesjournal.com

Guide for Authors International Journal of Advances in Applied Sciences (IJAAS) is a peer-reviewed and open access journal dedicated to publish significant research findings in the field of applied and theoretical sciences. The journal is designed to serve researchers, developers, professionals, graduate students and others interested in state-ofthe art research activities in applied science areas, which cover topics including: chemistry, physics, materials, nanoscience and nanotechnology, mathematics, statistics, geology and earth sciences. Papers are invited from anywhere in the world, and so authors are asked to ensure that sufficient context is provided for all readers to appreciate their contribution.

The Types of papers The Types of papers we publish. The types of papers that may be considered for inclusion are: 1) Original research; 2) Short communications and; 3) Review papers, which include meta-analysis and systematic review

How to submit your manuscript All manuscripts should be submitted online at http://iaescore.com/journals/index.php/IJAAS

General Guidelines 1) Use the IJAAS guide (http://iaescore.com/journals/ourfiles/IJAAS_guideforauthors_2019.docx) as template. 2) Ensure that each new paragraph is clearly indicated. Present tables and figure legends on separate pages at the end of the manuscript. 3) Number all pages consecutively. Manuscripts should also be spellchecked by the facility available in most good word-processing packages. 4) Extensive use of italics and emboldening within the text should be avoided. 5) Papers should be clear, precise and logical and should not normally exceed 3,000 words. 6) The Abstract should be informative and completely self-explanatory, provide a clear statement of the problem, the proposed approach or solution, and point out major findings and conclusions. The Abstract should be 150 to 250 words in length. The abstract should be written in the past tense. 7) The keyword list provides the opportunity to add keywords, used by the indexing and abstracting services, in addition to those already present in the title. Judicious use of keywords may increase the ease with which interested parties can locate our article. 8) The Introduction should provide a clear background, a clear statement of the problem, the relevant literature on the subject, the proposed approach or solution, and the new value of research which it is innovation. It should be understandable to colleagues from a broad range of scientific disciplines. 9) Explaining research chronological, including research design and research procedure. The description of the course of research should be supported references, so the explanation can be accepted scientifically. 10) Tables and Figures are presented center. 11) In the results and discussion section should be explained the results and at the same time is given the comprehensive discussion. 12) A good conclusion should provide a statement that what is expected, as stated in the "Introduction" section can ultimately result in "Results and Discussion" section, so there is compatibility. Moreover, it can also be added the prospect of the development of research results and application prospects of further studies into the next (based on the results and discussion). 13) References should be cited in text by numbering system (in Vancouver style), [1], [2] and so on. Only references cited in text should be listed at the end of the paper. One author should be designated as corresponding author and provide the following information: • E-mail address • Full postal address • Telephone and fax numbers Please note that any papers which fail to meet our requirements will be returned to the author for amendment. Only papers which are submitted in the correct style will be considered by the Editors.

International Journal of Advances in Applied Sciences (IJAAS) Institute of Advanced Engineering and Science (IAES) e-mail: ijaas@iaesjournal.com

IJAAS Journal Order Form Volume

Number

Amount

Price (Rp)

Price (USD)

Total

Name

Company

Address

City / State

Zip

Telephone/Fax

: ........................, ..........................

Signature: ..................................

Order form for subscription should be sent to the editorial office by fax or email

Payment by Bank Transfer Bank Account name (please be exact)/Beneficiary: LINA HANDAYANI Bank Name: CIMB NIAGA Bank Branch Office: Kusumanegara City: Yogyakarta Country :Indonesia Bank Account # : 5080104447117 Bank Swift : BNIAIDJAXXX >>> Please find the appropriate price in the price list on next page >>>

The price list for domestic and foreign subscribers

Volume

Number

Year

1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 8 8 8 8

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

2012 2012 2012 2012 2013 2013 2013 2013 2014 2014 2014 2014 2015 2015 2015 2015 2016 2016 2016 2016 2017 2017 2017 2017 2018 2018 2018 2018 2019 2019 2019 2019

Price (IDR) for domestic subscribers 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 290.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 319.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00 349.000,00

Price (USD) for foreign subscribers 36 36 36 36 36 36 36 36 40 40 40 40 40 40 40 40 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44 44

The price included the printing, handling, packaging and postal delivery fees of the hardcopy to the address of the authors or subscribers (with Registered Mail). For foreign subscribers, an additional fee is charged if you would your order is mailed via Express Mail Service (EMS): - $25 for ASIA Continent - $35 for AUSTRALIA Continent - $35 for AFRICA Continent - $39 for AMERICA Continent - $39 for EUROPE Continent (No additional fee for delivering your order by Registered Mail).

International Journal of Advances in Applied Sciences (IJAAS) Volume 7, issue 4, Dec. 2018

Articles inside

S. S. Tadasare, S. S. Pawar

S. Nagavalli, G. Ramachandran

F. Fareeza, Chunchu Rambabu, S. Krishnaveni, Abel Chernet Kabiso

Kaviyaraj. R, Karthika M S, Jeni Narayanan L.A, Saleekha

Siddhaling Urolagin, Ishita Upadhyaya, Afrah Abbas Thakur

Rajesh Keshavrao Deshmukh, Shilpa Markandey, Pooja Sahu

Nutenki Siddhartha, G. Renuka

Leila Belhadef, Zoulikha Mekkakia Maaza