Nuicone 2015)

2015 5th Nirma University International Conference on Engineering (NUiCONE 2015)

Ahmedabad, India 26-28 November 2015

IEEE Catalog Number: CFP1555R-POD ISBN: 978-1-4799-9992-7 LIST OF FULL PAPERS

ID No. Paper ID Title/Page No. Authors Analyzing effect of bad measurement data on load Jigar Patel, Daivat Desai, Vaibhav 38 555 flow and state estimation in power system 199 Patel, Dishang D. Trivedi and Santosh C. Vora VM Placement of Multidimensional Resources using Naisargi Patel and Govind Patel 39 350 Cartesian Co-ordinates Based Approach 205 Hierarchical Clustering Technique for Word Sense Nirali Patel, Bhargesh Patel, Rajvi 40 234 Disambiguation using Hindi WordNet 210 Parikh and Brijesh Bhatt Pose, Illumination and Expression Invariant Face Pradip Panchal, Palak Patel, 41 118 Recognition using Laplacian of Gaussian and Local Vandit Thakkar and Rachana Binary Pattern 215 Gupta Performance Enhancement of 12 X 160 Gbps (1.92 Rohit Patel and Dilip Kothari 42 519 Tbps) WDM Optical System for Transmission Distance upto 8000 km with Differential Coding 221 Arduino Controlled War Field Spy Robot using Jigneshkumar Patoliya, Haard 43 512 NightVision Wireless Camera and Android 227 Mehta and Hiteshkumar Patel Optimized Unscheduled Interchange Based Secondary Shital Pujara and Chetan Kotwal 44 310 Control for Two Area Deregulated Electricity Market 232 Introducing the Conceptual Model of Industrial Aarthi Raghavan 45 126 MOOCs (I-MOOCs) for Engineering Classes 240 Design Of Low Voltage Bandgap Reference Circuit Sushma Suresh Sangolli and 46 253 Using Subthreshold MOSFET 246 Rohini Hongal Design of Low Power CMOS Low Noise Amplifier Hardik Sathwara and Kehul Shah 47 590 Using Current Reuse Technique 252 H∞ Loop shaping technique based robust control Bhavin Shah, Gopinath Pillai and 48 585 design of SVC controller for power oscillations Pramod Agarwal damping considering global signal 258 A Distributed Dynamic and Customized Load Vedang Shah and Harshal Trivedi 49 507 Balancing Algorithm for Virtual Instances 264 Comparative analysis of zoning based methods for Ankit Sharma, Dipak Adhyaru, 50 84 Gujarati handwritten numeral recognition 270 Tanish Zaveri and Priyank Thakkar Optimal Placement of TCSC for Improvement of Aesha Sheth, Chetan Kotwal and 51 326 Static Voltage Stability 275 Shital Pujara Design and Implementation of Isolated 30 V, 30 A Anand Sheth, Vinod Patel, 52 415 DC Power Supply Using Synchronous Rectifier 281 Vijendra Kharbikar and Manisha Implementation of Edge Detection Algorithms in Real Ami Shukla, Vibha Patel and 53 503 Time on FPGA 288 Nagendra Gajjar EncryScation: A Novel Framework for Cloud IaaS, Krunal Suthar and Jayesh Patel 54 270 DaaS security using Encryption and Obfuscation Techniques 292 Design and implementation of BiCMOS based low Deepa Talewad, Anilkumar V 55 589 temperature coefficient bandgap reference using Nandi, Vaishail B M 130nm technology 297

Copyright and Reprint Permissions: Abstracting is permitted with credit to the source. Libraries are permitted to photocopy beyond the limit of U.S. copyright law for private use of patrons those articles in this volume that carry a code at the bottom of the first page, provided the per-copy fee indicated in the code is paid through Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923.

***This publication is a representation of what appears in the IEEE Digital Libraries. Some format issues inherent in the e-media version may also appear in this print version.

IEEE Catalog Number: CFP1555R-POD ISBN (Print-On-Demand): 978-1-4799-9992-7 ISBN (Online): 978-1-4799-9991-0 ISSN: 2375-1282

Additional Copies of This Publication Are Available From:

Curran Associates, Inc 57 Morehouse Lane Red Hook, NY 12571 USA Phone: (845) 758-0400 Fax: (845) 758-2633 E-mail: [email protected] Web: www.proceedings.com

2015 5th Nirma University International Conference on Engineering (NUiCONE)

Comparative analysis of zoning based methods for Gujarati handwritten numeral recognition

Ankit K. Sharma1, Dipak M. Adhyaru2, Tanish H. Zaveri3, Priyank B Thakkar4 1Assistant professor, 2 Professor & Head, 3Associate Professor, 4Associate Professor 1,2Instrumentation and Control Engineering Section, 3Electronics and Communication Engineering Section, 4Computer Engineering Section Institute of Technology, Nirma University, Ahmedabad, Gujarat, India [email protected], [email protected], [email protected], [email protected]

of the numerals. Whereas, printed numerals on the other hand, Abstract— Gujarati is one of the ancient Indian languages spoken are simpler to identify because of their uniformity. widely by the people of Gujarat state. This paper is concerned with the recognition of handwritten Gujarati numerals. For Numeral recognition can be done in two ways: Offline recognition of Gujarati numerals zoning based Feature recognition and online recognition. This paper emphasizes on extraction method is used. Numeral image is divided in 16x16, offline numeral recognition. Application of handwritten 8x8, 4x4 and 2x2 Zones. After feature extraction through the zoning method, Naive Bayes classifier and multilayer feed numeral recognition includes, helping sort out or categorize forward neural network classifier are implemented for the postal mails, bank cheque, code reading, postal address classification of numerals. For the database generation, 14,000 reading, form processing, signature verification, etc. Optical samples of each numeral are used. The overall recognition rates character recognition helps to reduce human efforts of of this method used for recognition of Gujarati numeral using manually handling and processing documents. 16x16, 8x8, 4x4 and 2x2 zoning with neural network are 93.03%, 95.92%, 91.89% and 61.78% and with Naive Bayes classifier are II. REVIEW OF RELATED WORK 75%, 85.60%, 81% and 53.75% respectively. In comparison to other foreign languages like Chinese, Index Terms— Gujarati script, Neural networks, Naive Bayes English, and Japanese etc., not much work has been carried classifier, Zone based feature extraction. out in the area of Gujarati numeral recognition. It is found that in comparison with Indian languages like Bangla, Hindi, I. INTRODUCTION Marathi, the OCR activities related to Gujarati language is Handwritten numeral Recognition becomes a prime area of very less. Shailesh A. Chaudhari and Ravi M. Gulati have research because of its potential, that can be used in various worked on separation and identification of mixed English – applications. Researchers have explored many Indian Gujarati printed numerals. Statistical approach is used as languages such as Hindi, Marathi, Telugu, Bangla, feature extraction with KNN classifier. An overall accuracy of Gurumukhi, Tamil and Kannada etc., but Gujarati is yet to be 99.23% is obtained with the same and an accuracy of 99.26% researched, explored and recognized to be on part with the for Gujarati and 99.20% for English numerals is obtained other Languages. This paper throws light on the Gujarati using KNN classifier [25]. Baheti M.J and Kale K.V have numeral recognition. Gujarati script is part of the Brahmic developed an algorithm to classify handwritten Gujarati family and it is similar to Devanagari script as most of the numerals using affine invariant moment feature extraction words are derived from Sanskrit. There is no header line at the technique. Authors have used various classifier to classify top of the letters or words in Gujarati script. Gujarati is the Gujarati numerals and highest accuracy of 92.28% is achieved mother tongue of the people of Gujarat state, one of the most using support vector machine. Accuracy of 90.04%,87.2% and spoken native languages and nearly 65 million people 84.1% is obtained using K-Nearest Neighbor, Gaussian converse in Gujarati. Gujarati numerals have various shapes distribution function and principal component analysis and many numerals have close resemblance which creates classifier [26]. In [27] affine invariant moments based feature confusion and have possibilities of incorrect recognition. This is used by Mamta maloo and K.V. Kale to classify paper talks about Optical character recognition for handwritten handwritten Gujarati numerals. Recognition rate of 91% is Gujarati numerals. Optical character recognition, usually obtained using support vector machine classifier.To recognize abbreviated as OCR,it translates typewritten, scanned copy of handwritten Gujarati numerals author Avani R. Vasant, images or documents into machine encoded text. This Sandeep R.Vasant and Dr. G.R.Kulkarni have used Neural translated machine encoded text can be easily searched, edited Network classifier. Recognition rate of 87.29%, 88.52% and and processed in numbers of ways as per our requirement. 88.76% is obtained for 7x5, 14x10 and 16x16 size images Handwritten document recognition is a very challenging area respectively [28]. In [8] Minimum hamming distance classifier for research and many researchers are putting in efforts to and k-NN classifier are used for identification of Gujarati convert handwritten scripts to computer readable format. characters and overall accuracy achieved was 67%. An Handwritten numeral recognition is tedious because algorithm for Gujarati script identification is proposed by S. handwritten numerals may vary from person to person K. shah and A.Sharma. Algorithm uses template matching depending on their writing style, curve, thickness and the size based approach for character classification [13]. Jignesh Dholakia, Atul Negi and S. Ram Mohan have used GRNN and k-NN classifier for classification of printed Gujarati characters