Proceedings of the 2009 International Conference on Signals ...

Proceedings of the 2009 International Conference on

Proceedings of the 2009 International Conference onSignals, Systems and Automation


Prof. Chintan ModiProf. Rahul Kher


(ICSSA 2009)

Editor in Chief

Dr. Himanshu Soni

Prof. Chintan ModiProf. Rahul Kher


(ICSSA 2009)

Editor in Chief

Dr. Himanshu Soni

Editors

Prof. Chintan Modi Prof. Hitesh ShahProf. Rahul Kher Mr. Nilesh Desai


(ICSSA 2009)

Editor in Chief

Dr. Himanshu Soni

Prof. Hitesh ShahMr. Nilesh Desai


Prof. Hitesh Shah Mr. Nilesh Desai


Proceedings of the 2009 International Conference on

Proceedings of the 2009 International Conference on Signals, Systems and Automation (ICSSA 2009)

Copyright © 2010 Himanshu Soni All rights reserved.

No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system,

without written permission from the publisher

Universal-Publishers Boca Raton, Florida

USA • 2010

ISBN-10: 1-59942-869-5 ISBN-13: 978-1-59942-869-7

www.universal-publishers.com

International Conference on Signals, Systems & Automation - 2009

iii

About Charutar Vidya Mandal

DR. C. L. PATEL

CHAIRMAN, CVM CHARUTAR VIDYA MANDAL

ESTD: 1945

One of the greatest son of India Late Sardar Vallabhbhai Patel inspired Shri Bhaikaka and Shri Bhikabhai Saheb for rural resurgence of post-independent India through education, and Charutar Vidya Mandal was born. Charutar Vidya Mandal was established in the year 1945 as a charitable trust with a prime objective of rural development through education to bring about the social awakening, social upliftment and enrichment. The uniqueness of Charutar Vidya Mandal lies in its ability to use quality education as a powerful means of social transformation. It was a stupendous task for the founders to establish a visionary organization; but the large-heartedness and high sense of philanthropy of this region made this possible. Over the subsequent years, Dr. H M Patel consolidated the efforts put in by the founders. Later on, in the 1990s, when Dr. C L Patel took over the reigns of Charutar Vidya Mandal as the Chairman, the country was facing a major economic and ideological change paving the way for globalization and liberalization. The dynamic leadership, missionary zeal and visionary outlook of Dr. C L Patel successfully took up the challenges. Various self-financed educational institutions started being established in the areas of Technology, Science and Engineering, Commerce and Management, offering emerging and innovative courses and programs such as Mechatronics, Automobile Engineering, Bio-technology, Food Processing Technology, E-Commerce, Valuation, etc. Today Charutar Vidya Mandal is empowering budding graduates to live up to the ever-changing environment and equipping them to face the Third Millennium with confidence and competence. Building competitive advantages is the renewed focus of Charutar Vidya Mandal in the 21st century. At present, Charutar Vidya Mandal operates 30 Educational Institutions from schools to colleges, and a sophisticated Research Institute, with over 30,000 students on the rolls.

About G H Patel College of Engineering & Technology

The institute is managed by one of the highly reputed and largest educational trust Charutar Vidya Mandal. From the very inception (in the year 1996) GCET has striven to develop itself into an ‘Institution of Excellence’ in education and research in consonance with the spirit of modern Gujarat. In meeting this challenge for excellence, GCET has shaped its institution to the contemporary as well as future demands of need oriented education. Today, GCET has total strength of 1800 students studying for Chemical, Mechanical, Information Technology, Electronics and Communication, Computer, Electrical Engineering and Mechatronics under tutelage of 116 dedicated and competent faculties. GCET is one of the premier Self-Financed Institutes in the state. GCET is the pioneer in offering Information Technology and Mechatronics Engineering programmes. It is equipped with state of the art computer laboratories with more than 450 nos. of P4 computers with 10Mbps shared internet Wi-Fi connectivity. More than 200 international journals are available in fully automated digital library. 100% placement of eligible students are secured regularly in various Professional and Multinational companies like Infosys, Wipro, TCS, L&T, Essar, MBT, Relience, PCS, Accenture, Siemens, DLF, Voltas, Tata Chemicals etc. GCET is ranked among top 100 engineering colleges across the country by Outlook ‘06 & ‘09.


iv

Our Vision

To produce engineering graduates who are globally competitive, live by a set of core values, are able to accept any

professional challenge thrown at them, and remain responsive to the needs of India and the humanity.

Our Mission

To foster a stimulating learning environment, develop excellence amongst students, faculty and staff in every activity

GCET carries out, and thereby aim to become one of the best premier technical institutes of the country.

Our Quality Policy

We, at GCET, will continuously strive to become and remain leaders amongst technical education institutions in India

through constant improvement in teaching learning process, continuous interaction with industry through consultancy,

combined development project and providing project and providing an intellectual environment conductive to lifelong

learning.

About Department of Electronics & Communication Engineering

The Department of Electronics & Communications started in 1997 is one of the largest and upcoming departments of G H Patel College of Engg. and Technology. This Department has given its first meritorious batch of EC graduates in July 2001; majority students of this batch are well placed through campus or have gone for higher studies. The Department has excellent laboratory facilities both for teaching and research. In 2007 the department started Post graduate program in Communication Engineering. The first batch of PG graduated in July 2009. In a short span of 3 years the department has published many international Journal and conference papers. The Department organizes workshops, seminars in various specialties, and runs summer/winter schools for the benefit of engineering college teachers and professional engineers.


v

Organizing Committee

Chief Patron Dr. C. L. Patel

Chairman, Charutar Vidya Mandal, Vallabh Vidyanagar, India

Patron Shri. R. P. Patel

Hon. Secretary, Charutar Vidya Mandal, Vallabh Vidyanagar, India

President Dr. Anurag Verma

Principal, G H Patel College of Engineering & Technology, Vallabh Vidyanagar, India

Chief Advisor Prof. J.C. Panchal

Advisor, EC & EE Department, G H Patel College of Engineering & Technology, Vallabh Vidyanagar, India

Convener Dr. H. B. Soni

Professor & Head, Department of Electronics & Communication Engineering,

G H Patel College of Engineering & Technology, Vallabh Vidyanagar, India

Organizing Secretary Prof. C. K. Modi

Professor, Department of Electronics & Communication Engineering, G H Patel College of Engineering & Technology, Vallabh Vidyanagar, India

Joint Organizing Secretaries

Prof. H.B.Shah, Prof. R. K. Kher Professors, Department of Electronics & Communication Engineering,

G H Patel College of Engineering & Technology, Vallabh Vidyanagar, India


vi

International Advisory Committee

Dr. U. B. Desai (Director, Indian Institute of Technology, Hyderabad) Dr. H. G. Ryu (Chungbuk National University, Republic of Korea)

Mr. Santosh Patel (CEO, Skyview Communication, USA) Mr. Vishal Gandhi (CEO, DigiDMS, USA) Mr. Amrish Patel (CEO, Piqquel Inc, USA)

Dr. M. Gopal (Professor, Indian Institute of Technology, Delhi) Dr. S. Mahapatra (Professor, Indian Institute of Technology, Bombay)

Dr. S. N. Merchant (Professor, Indian Institute of Technology, Bombay) Dr. P. S. V. Nataraj (Professor, Indian Institute of Technology, Bombay)

Dr. Vinod Kumar (Indian Institute of Technology, Roorkee) Dr. R. S. Anand (Indian Institute of Technology, Roorkee)

Prof. O.P.N. Calla (Director, International Center for Radio Science, Jodhpur) Dr. N. D. Jotwani (Director, School of Petroleum Technology, Gandhinagar)

Dr. F. S. Umrigar (Principal, BVM Engineering College, Vallabh Vidyanagar) Dr. R. K. Jain (Principal, ADIT Engineering College, New Vallabh Vidyanagar)

Dr. D. N. Bhatt (Director, Institute of Science & Technology for Advanced Research, Vallabh Vidyanagar) Mr. Nirbhay K Chaubey (Treasurer , IEEE Gujarat Section)

Mr. Kashyap Mankad (Scientist, Space Application Center, ISRO, Ahmedabad) Mr. Anup Shah (CEO, Insignex, Anand)

Technical Programme Steering Committee

Dr. V. N. Kamat (Director, Center for Apparent Energy Research, Vallabh Vidyanagar) Prof. Indrajit Patel (B & B Institute of Technology,Vallabh Vidyanagar)

Dr. C. D. Parikh (Professor, DA-IICT, Gandhinagar) Dr. M. V. Joshi (Professor, DA-IICT, Gandhinagar)

Dr. M. A. Zaveri (Professor, National Institute of Technology, Surat) Dr. P. Bharath Kumar (Taiwan Semiconductor Manufacturing Corporation, Taiwan)

Dr. G. Kannan (G. M. India Science Research Lab., Bangalore) Dr. Manoj C. R (Professor, National Institute of Technology, Callicut)

Dr. Jayalaxmi (Patni Computer Systems Ltd., Mumbai) Dr. Bhushan Jagyasi (Tata Consultancy Services Ltd., Mumbai)

Dr. V. K. Thakar (Head, ADIT Engineering College, New Vallabh Vidyanagar) Dr. T. D. Pawar (Professor, BVM Engineering College, Vallabh Vidyanagar)

Mr. Gautam Kappila (Texas Instruments, Bangalore) Mr. Sameer Ranpara (Intel, USA)

Prof. A. M. Shah (Head CP Dept, GCET Engineering College, Vallabh Vidyanagar) Prof. M. S. Patel (Head, IT Dept, GCET Engineering College, Vallabh Vidyanagar)

Prof. P. B. Swadas (Head, CP Dept., BVM Engineering College, Vallabh Vidyanagar) Prof. U. S. Deopurkar (Head, EC Dept., BVM Engineering College, Vallabh Vidyanagar)

Mr. Nirbhay K Chaubey (Prof.,CP Dept.,ISTAR, Vallabh Vidyanagar Treasurer,IEEE Gujarat Section) Mr. Sachin Vyas (Sonim, Banglore)

Mr. Keyur Mistry (Texas Instruments, Bangalore) Mr. Mahesh Shah (GE, Bangalore)

Mr. Snehal Desai (Larsen & Tubro, Baroda)


vii

Mr. Aditya Tatu (University of Denmark) Mr. Abhishek Bhadauria (Alcatel-Lucent, Bangalore)

Mr. A. S. Bishnoi (Asst. Director, MSQAA, DRDO, Hyderabad) Dr. Rajkiran Panugant ( Microsoft , USA)

Dr. Kriyang Shah ( La Trobe University, Australia) Dr. C.V.Jiji ( Government College, Kannur, Kerala)

Mr. Arnav Bhavsar (Indian Institute of Technology, Madras) Mr. Rushiraj Rathod (Tata Consultancy Services, Pune)

Mr. Jay Shah (Qualcomm, USA) Prof. N J Kothari (DDIT, Nadiad, India)

International Program Committee (Reviewers)

Dr. Vithal Kamath (CEO, Baroda Electric Meters) Prof. Bhaskar Thakkar (EC Dept, GCET Engineering College, India)

Dr. Tanmay Pawar (EC Dept, BVM Engineering College, India) Mr. Paresh Patel (CEO, SLS Inc., USA)

Prof. Vijay Makwana (EE Dept, GCET Engineering College, India) Mr. Nitin Paranjape (CEO, Edutech Systems, India)

Mr. Santosh Patel (CEO, Skyview Communication Inc., USA) Prof. Chintan K. Modi (EC Dept, GCET Engineering College, India)

Prof. N P Gajjar (EC Dept, Nirma University, India) Prof. Rahul Kher (EC Dept, GCET Engineering College, India)

Dr. Bhushan Jagiyasi (TCS, India) Mr. Chandrakant Padole (Mumbai, India)

Ms. Dipthi Chander (Indian Institute of Technology, Bombay, India) Prof. Himanshu Soni (EC Dept, GCET Engineering College, India)

Dr. Manjunath Joshi (Professor, DAIICT, Gandhinagar, India) Prof. Mehul Shah (EC Dept, GCET Engineering College, India)

Dr. Rajkiran Panuganti (Microsoft, USA) Prof. Ramji Makawana (EC Dept. ADIT Engineering College, India)

Dr. Aditya Tatu (University of Denmark) Prof. N M Patel (EC Dept, BVM Engineering College, India)

Dr. Mukesh Zaveri (Professor, National Institute of Technology, Surat, India) Prof. Hitesh Shah (EC Dept, GCET Engineering College, India) Dr. V K Thakar (EC Dept, ADIT Engineering College, India)

Prof. Malay Bhatt (CP Dept, DDIT Engineering College, India) Mr. Anup Shah (CEO, Insignex, India)

Dr. C. V. Jiji (Kerala, India) Dr. R S Anand (Indian Institute of Technology, Roorkee, India)

Dr. Vinod Kumar (Indian Institute of Technology, Roorkee, India) Dr. Kriyang Shah (LaTrobe Uni, Australia)

Prof. Sumil Kavam (Indian Institute of Technology, Mumbai, India) Prof. Anand Darji (National Institute of Technology, Surat, India)

Prof. Rajesh Thakkar (Govt. College Gandhinagar, India) Dr. Deepak Fulwani (EE Dept, GCET Engineering College, India)

Dr. Axay Mehta (EE Dept, GCET Engineering College, India)


viii

Mr. Mukesh Bhesania (EE Dept, GCET Engineering College, India) Mr. Ritesh Patel (EE Dept, GCET Engineering College, India)

Dr. Bhavesh Bhalja (EE Dept, ADIT Engineering College, India) Mr. Pragnesh Shah (EE Dept, ADIT Engineering College, India)

Mr. Dhaval Patel (Indian Institute of Technology, Mumbai, India) Dr. S. Gopinath (ABB, Banglore, India) Mr. Arunkumar (ABB, Banglore, India) Mr. Dipak Adhyaru (Nirma Uni., India)

Mr. Anand Patel (Ganpat Uni. India) Prof. Apuva Shah (CP Dept, GCET Engineering College, India)

Local Organizing Committee

Mr. Nilesh Desai Mr. Pradeep Shah Ms. Geetali Dutta Mr. Hitesh Loriya

Mr. Navin Ganeshan Mr. Samir Trapasiya

Mr. Hasmukh Koringa Mr. Kavindra Jain

Mr. Aslam Durvesh Mr. Falgun Thakkar Mr. Mayank Mahant Mr. Ahish Christian

Mr. Nirav Desai Mr. Giriraj Patel Ms. Smita Joshi

Ms. Kinjal Mehta Ms. Foram Shelat

Ms. Parita Rao Mr. Kalpesh Patel Mr. Rakesh Patel

Mr. Kaushik Sargara Prof. Yogesh Chauhan Prof. Kumarpal Trivedi Prof. Nikhil Gondaliya

Mr. Deven Agravat Ms. Jasmin James

Mr. Yamnesh Khamar Mr. Anupam Patel


ix

Table of Contents

Paper for Oral Presentation

Image Processing

• Medical Image Analysis: Recent Trends and Future Scopes Heena Kher, Rahul Kher & Dr. Tanmay Pawar, Gujarat, India…………………………………………………..

1

• Quality Evaluation of Cuminum Cyminum L (Cumin) Seeds Using Colorization Jalpa J Patel, Chintan K Modi & Kavindra R Jain, Vallabh Vidyanagar, Gujarat, India……………………

7

• An Algorithm for Language Independent Speaker Identification Amit Kaul & Sushil Chauhan, Hamirpur, Himachal Pradesh, India Manmohan Singh & A. S. Arora, Longowal, Sangrur, Punjab, India……………………………………………..

13

• Signal processing for control room EMI analysis using frequency domain approach Vidya. R. Keshwani, Prof. Naveeta Kant & Dr. Kallol Roy, Mumbai, Maharashtra, India……………………

18

• Knowledge based Extraction of Fetal Heart Rate from Phonocardiographic Signals Vijay S. Chourasia & A. K. Mittra, Gondia, Maharashtra, India………………………………………………….

24

• Galois field based error detection and correction, auto finding and decoding of two dimensional barcode from an image for automatic label inspection system

Rakesh V.Gosai, Navin Mandal, Shital P. Thakkar, Gujarat, India……………………………………………….

29

• Machine Vision Based Liquid Level Inspection System using Laplacian of Gaussian Edge detection Technique

Ishan Doshi, Galav Dani, Bhaumit Patel, Vedang Chauhan & Chintan K. Modi, Gujarat, India…………….

35

• License Plate Extraction Based on Gradient and Wavelet Analysis Chirag N. Paunwala & Dr. Suprava Patnaiak, Surat, Gujarat, India…………………………………………….

41

• An Adaptive Efficient Blind Algorithm for Digital Watermarking Mita C. Paunwala & Dr. Suprava Patnaiak, Surat, Gujarat, India……………………………………………….

45

• Optical Character Recognization of Gujarati Numericals Hinal Shah & Anup Shah, Gujarat, India……………………………………………………………………………..

49

• Self Quotient Image and Multi-Scale Filtering for Illumination Invariant Face Recognition Ramji M. Makwana & Vishvjit K. Thakar, Gujarat, India………………………………………………………….

54

• Performance Evaluation of Skew Detection Algorithms for Gujarati Text Images Shital P. Thakkar & Amit H. Choksi, Gujarat, India………………………………………………………………...

60

• Image Registration in Time and Frequency Domain for Recovering Affine Transformation Mehfuza Holia & Vishvjit K. Thakar, Gujarat, India………………………………………………………………..

65

• Comparative study of different quantization techniques for natural and medical image compression Falgun Thakkar, Rahul Kher, Chintan Modi & Heena Kher, Gujarat, India…………………………………….

71

• Content Based Image Retrieval using different Colormaps Kalpesh M. Shah & Vishvjit K. Thakar, Gujarat, India……………………………………………………………..

77

• Morphological Filter: Preprocessing Step in Edge Detection for Noisy Grayscale Images Mehul Thakkar, Shailesh Khant & Hitesh Shah, Gujarat, India…………………………………………………

81


x

Digital Signal Processing

• Line Identification in the Presence of Colored Noise Amir H. Khanshan, Islamic Azad University, Kahsan branch, Iran………………………………………………

• An Efficient Reconfigurable Architecture for Fractional Fourier Transforms Soumen Mukherjee & Anal Acharya, India…………………………………………………………………………

85

• Diagnosis Of Diabetes Mellitus Using Artificial Neural Network Based E-Nose For Telemedicine Maria Jamal, M R Khan, S A Imam & Arif Jamal, Delhi, India…………………………………………………...

89

• Text-To-Speech Synthesis for Gujrathi Language Prof. Vaishali R. Wadhe & Prof. Sandeep S. Sahare, Mumbai, India…………………………………………….

95

• Performance Evaluation of LMS & NLMS Algorithms for Noise Minimization from Speech Signals V. K. Gupta & Mahesh Chandra, Ranchi, India S. N Sharan, Greater Noida, India. Omar Farooq, Aligharh, India………………………………………………

100

Embedded Systems

• Three-Phase Power Quality Monitoring System Ankit Shah, Gujarat, India………………………………………………………………………………………………

104

• Implementation of Embedded Intelligent Control System Hatkar A. P. & Chougule D. G., Kolhapur, India…………………………………………………………………

110

• Embedded Computer for Real Time Monitoring of Patients Mr. T. M. Pattewar & Prof. Ms. R. W. Jasutkar, Nagpur, India…………………………………………………

115

• Implementation of Semaphore: A Solution to the Priority Inheritance Problem in uc/OS Real-Time Kernel Tareek M Pattewar & Nitin N. Patil, Shirpur, India………………………………………………………………

119

• Design and Implementation of Elevator Controller Using Embedded Tool Altium Designer Ilesh M Parmar & Nilay N Bhuptani, Gujarat, India………………………………………………………………

122

• Wrist Pulse Classifier Utilizing Frequency Domain Features Bhaskar Thakker & Anoop Lal Vyas, Delhi, India…………………………………………………………………..

127

• Microcontroller based Programmable Electronic Bell System Bhaskar Thakker & Jigisha Thakker, Gujarat, India………………………………………………………………

132

• A Single-Sensor, Chip Based Solution For Monitoring Body Temperature In Real Time For Wearable Health Monitoring Dipak Patel, Tanmay Pawar, Kaushika Patel & Arjun Bhamaniya, Gujarat, India……………………………

136

• Digital Blood Pressure Meter Implemented using Embedded and LABVIEW Parul Panchal, Kaushika Patel, Hiren Patel, Gujarat, India………………………………………………………

139

• Embedded Impedance Sensor Design for Future Lunar Wireless Sensor Network with ‘Chirp’ Type Perturbation Signal J. P. Pabari & Y. B. Acharya, Ahemdabad, Gujarat, India U. B. Desai & S. N. Merchant, Mumbai, Maharashtra, India……………………………………………………..

144


xi

RF, Microwave & Satellite Communication

• High Radiation Performance of a Rectangular Microstrip Patch Antenna with Air Substrate Nagendra Kumar, Chandan Kumar, Uday Shanker Kanth & Sudipta Chattopadhyay,West Bengal, India…

148

• Design, Simulation and Test of a Microstrip Patch Antenna Array Twinkle V. Doshi & Khyati R.Zalawadia, Gujarat, India…………………………………………………………..

151

• Study on Antenna Pointing Errors for Multibeam GEO Satellites Rajesh Kumar Singh & Neha Chadha, Gujarat, India……………………………………………………………

156

Wireless Communication

• Achieving Full Frequency and Space Diversity in OFDM and STBC in Wireless System Vigyan Agrawal & Ashutosh Sharma, Madhya Pradesh, India……………………………………………………

160

• Factors influencing the Wireless Sensor Network Design Sachin Gajjar, Gujarat, India…………………………………………………………………………………………

166

• Comparative Analysis of Slow selective-fading channel using Space-Time Coding for Next Generation Wireless Communication

Kaushal P. Makhecha, Sunil B. Bhatt, Megha Mehta, K. H. Wandra, Gujarat, India………………………….

170

• The WiMAX Physical layer modeling for image processing application Kaushal P. Makhecha, Sandip J. Dawda & Anil C. Suthar, Gujarat, India…………………………………….

176

• SLM and Modified SLM Scheme for PAPR Reduction of OFDM System Ms Khyati K. Desai & Ms Jigisha N. Patel, Gujarat, India………………………………………………………..

181

• FPGA Implementation Of OFDM Modem Swapnil A. Shah & Jigisha N. Patel, Gujarat, India………………………………………………………………...

186

• A Survey: MAC Misbehaviour in Wireless Mesh Networks Anil Kumar Gankotiya, Sahil Seth & Gurdit Singh, Chandigarh, India………………………………………….

192

• Comparison of Hard Output and Soft Output Viterbi Algorithm Decoding Techniques Mrs. D. M. Khatri, Prof. S. L. Haridas & Dr. N. K. Choudhari, Wardha, Maharashtra, India……………….

197

• Coded Orthogonal Frequency Division Multiplexing: Performance over Frequency Selective Channel Hardip Shah & Chandresh Parekh, Gujarat, India…………………………………………………………………

202

• Comparison of Simulated Annealing Approach and A Novel Tracking Stragtegy to Reporting Cell Planning Problem of Mobile Location Management Falguni Mehta & Prashant Swadas, Gujarat, India………………………………………………………………...

208

• Performance Enhancement of TCP New Reno Kaushika Patel, Paru; Panchal & Hardik Patel, Gujarat, India……………………………………………….....

213

• Comparative study of different Multiuser detection techniques for DS-UWB based system Durvesh Aslam, Prof. Himanshu Soni & Mayank Mahant, Gujarat, India………………………………………

219

• Performance Evaluation of GSM Network Using Qualnet Simulator Mr. M. R. Khandhedia, Mr. S. B. Bhatt, Ms. V.B.Kanani, Prof. D. N. Khandhar, Prof.K. H. Wandra, India.

224

• Design of Viterbi Decoder using Minimum Transition Register Exchange Method Prof. S. L. Haridas & Dr. N. K. Choudhari, India…………………………………………………………………..

227

• Compensation schemes and Performance Analysis of IQ imbalances in Direct Conversion Receivers Anjali Pise & Achala Deshmukh, Pune, Maharashtra, India………………………………………………………

231


xii

VLSI

• Offset Reduction in a Low Voltage, Low Power current feedback amplifier Amisha P. Naik & N. M. Devashrayee, Gujarat, India……………………………………………………………..

237

• Design and Simulation of Ternary Logic Based Arithmetic Circuits Kanchan S. Gorde and Dr. P. R. Deshmukh, India…………………………………………………………………

242

• Design and Performance Evaluation of FFT Algorithms on FPGA-based Multi-Processor systems Jishnu Dave, Parth Lakhiya & Prof. Nagendra Gajjar, Gujarat, India…………………………………………..

247

• Effect of Scaling on Two Stage CMOS Op-Amp Umar Faruque Khan & Dr. S. C. Bose, India………………………………………………………………………..

252

• Implementation of Digital PID Controller Using FPGA Manan A Modi & Rakesh C Patel, Gujarat, India…………………………………………………………………..

260

• An 8-Bit Pipeline ADC Manish I. Patel & N.M.Devashrayee, Ahemdabad, Gujart, India…………………………………………………

264

Advanced Control • 2-DOF Controller Synthesis for Tracking of Harmonic Reference Trajectories

Markana Anilkuma, Nishant Parikh & Ankit Shah, Maharashtra, India………………………………………… 269

• Design of Digital Controller Using Pole Placement Approach For Magnetically Suspended Ball System

Sharad P Jadhav, C.B Kadu, Maharashtra, India…………………………………………………………………

275

• Analysis, Simulation and Implementation of Wide Line Voltage Range, Two Switch, Power Factor Corrected AC-DC Buck Converter Mahadev S. Patil, Devendra N. Kyatanavar & Sanjay P. Patil, Maharashtra, India………………………….

280

• Iterative learning control of spatio-temporal dynamics with piece-wise constant actuation Blazej Cichy & Krzysztof Galkowski, Poland Eric Rogers, United Kingdom…………………………………………………………………………………………..

285

• Design of Microcontroller Based Data Acquisition System with Ethernet Communication Trupti P. Agarkar, Sonali N. Kulkarni, Mukesh D Patil, Tahir Kamal Khan & R. Balasubramanian, India.

290

• Robust Tracking Control of Unmanned Aerial Vehicle with Sliding Mode Control Systems: A Multirate Output Feedback Approach Brijesh Naik, Surat, India……………………………………………………………………………………………….

296

• Optimal Control of Three-Contact Planar Manipulation S. K. Sharma, S. Mukherjee & I.N. Kar, New Delhi, India…………………………………………………………

302

• Innovative Tank Management System by Distributed Control System Amir Firoozshahi & Foroozan Siyahpoush, Tehran, Iran………………………………………………………….

307

• Heuristic Algorithms for Packet Classification Mrs. Mrudul A. Dixit & Dr. B. V. Barbadekar, Maharashtra, India…………………………………………….

313

• Performance Issues of Smart Card based on line Health Care Automation System Narendra Kohli & Nishchal K.Verma, Kanpur, India………………………………………………………………

317

• Repetitive Control using 2D System Theory S. Gopinath, I. N. Kar & R. K. P. Bhat, India………………………………………………………………………..

323


xiii

Other

• Fault Measuring Technique for Neural Hardware Amit Prakash Singh, Pravin Chandra & Chandra Sekhar Rai, Delhi, India…………………………………….

329

• Cryptography using Chaotic Neural Network Sukant Kumar Chhotaray & Girija Sankar Rath, India…………………………………………………………….

334

• Least phase angle difference method for PD location in transformer windings V.Jeyabalan & S.Usa, Chennai, India…………………………………………………………………………………

340

• Ambulation Study of ECG in Wearable Devices Dipak Vala, Rahul Kher, Dr. Tanmay Pawar & Dr. V. K. Thakar, Gujarat, India……………………………

344

Medical Image Analysis: Recent Trends and Future Scopes

Rahul Kher

Asst. Professor, EC Dept, GCET,

Vallabh Vidyanagar. [email protected]

Dr. Tanmay Pawar Asst. Professor, EL Dept,

BVM Engg. College, Vallabh Vidyanagar.

[email protected]

Heena R. Kher Lecturer, EC Dept,

ADIT, New Vallabh Vidyanagar. [email protected]

Abstract- In recent times, a lot of medical imaging modalities and related technologies have been invented with a sole goal of upliftment of human being in terms of better health and living. These modalities include computed tomography (CT), magnetic resonance imaging (MRI), X-ray, ultrasound, positron emission tomography (PET), single photon emission computed tomography (SPECT), and fluorescence microscopy. The efficient use of these modalities and attached technologies such as picture archival and communication (PACS); teleradiology and telemedicine; digital imaging and communication in medicine (DICOM) etc. are extremely useful in diagnosing the anatomical abnormalities in the human beings. A wide range of techniques have been devised and the efforts are still going on to enhance the quality of these images. Apart from enhancement, several other techniques such as compression, segmentation, speckle-noise reduction, retrieval, registration, fusion etc have been invented and explored. This paper focuses on the recent trends of such medical image analysis techniques and technologies and the possible scope for the future research.

Keywords: Medical imaging, CT, MRI, PACS, DICOM, Compression, Segmentation

I. INTRODUCTION Modern medical imaging devices, such as computed

tomography (CT), MRI, and electronic endoscopy, provide tremendous benefits for easy disease diagnoses. Corresponding computer technologies play important roles in processing and analyzing medical images, including computer graphics, pattern recognition, virtual reality, etc. Three main research fields on which medical image processing and analyzing focuses are structural imaging, functional imaging, and molecular imaging. Past two decades have witnessed many algorithms developed in these fields by scientists and engineers, and recently, new algorithms have been emerging continuously [1]–[3]. The area is very challenging which has enormous utility. A wide range of image analysis methods viz. enhancement, noise-reduction, segmentation, compression, retrieval & registration, fusion etc have been implemented in the recent times. Not only that but the interfaces two or more domains like information technology in

biomedicine, medical expert system [4],[5], mobile in medicinal diagnosis etc have also emerged as the current research areas. The current research trends in these methods and technologies have been discussed in sections II to V followed by the conclusions and further scopes in section VI.

II. COMPRESSION Modern medical imaging requires storage of large quantities

of digitized clinical data. Due to the constraints of bandwidth and storage capacity, however, a medical image must be compressed before storage and transmission. It is of utmost importance to preserve the quality of a medical image after compression. Many lossless schemes have therefore been proposed.

A. Perceptually Lossless Medical Image Coding Wu et al. have suggested a perceptually lossless medical

image coding technique [6]. Based on the JPEG2000 coding framework [7], the heart of the perceptual coder (PC) is the implementation of an advanced visual pruning function combined with a human vision model [8], [9] to identify and to remove visually insignificant/irrelevant information as well as to offer the benefits of simplicity and modularity. Furthermore, the visual pruning function can be embedded into any discrete Wavelet transform-based coder while maintaining bitstream compliance. This has been demonstrated previously in [10], [11] based on the Set Partitioning of Hierarchical Trees (SPIHT) coding framework [12].

B. ROI Coding Techniques for Medical Image Compression • Current compression schemes produce high

compression rates if loss of quality is affordable. However, in most cases physicians may not afford any deficiency in diagnostically important regions of images; called regions of interest (ROIs). An approach that brings a high compression rate with good quality in the ROI is thus necessary. The general theme is to preserve quality in diagnostically critical regions while


Page 1 of 348

allowing lossy encoding of the other regions (see Fig. 1). The aim of the research focused on ROI coding is to allow the use of multiple and arbitrarily shaped ROIs within images, with arbitrary weights describing the degree of importance for each ROI including the background (i.e., image regions not belonging to ROI) so that the latter regions may be represented by different quality levels. In this context, the ROI coding techniques can be classified according to the image type they apply to; thus the first class includes ROI coding schemes developed for two-dimensional (2-D) still medical images whereas the second class consists of ROI coding in the case of volumetric images. In the third class, a prototype ROI encoder for compression of angiogram video sequences has been presented [13].

• In another ROI based coding scheme, called the contextual coding scheme, proposed by Ansari et al., the basic objective is to achieve high compression rates by applying different compression thresholds for the wavelet coefficients of each DWT band (BG & ROI) while conventional image compression methodologies utilizing the DWT apply it to the whole image. That is, in the contextual coding, different compression rates are applied to the wavelet coefficients in different ROls respectively resulting in high CRs and diagnostic quality of the image with sufficient information retained in the background. Further, contextual coding provides an excellent trade-off between reconstructed image quality and compression ratio. In order to find an optimal technique for medical image compression, an experimental study is conducted to qualitatively judge the efficacy of contextual approach in comparison with the EBCOT, Maxshift and Implicit compression techniques applied to medical images and it is found that contextual SPIHT (CSPIHT) reconstructed images outperform the above methods in terms of performance and the visual quality [14].

C. Medical Image Compression on Mobile Devices Medical applications have already been integrated with

mobile devices and are being used by the medical personnel in treatment centers, for retrieving and examining patient data and medical images. In [15] Doukas et al. have suggested an application designed for viewing DICOM compliant medical images using wavelet compression on mobile devices. The proposed application uses an image compression algorithm called distortion limited wavelet image codec (DLWIC) introduced by J. Lehtinen in [16]. The specific algorithm solves the problem of distortion limiting (DL) by allowing the user of the algorithm to specify the mean square error (MSE) of the decompressed image as controlling parameter for the compression algorithm. Furthermore, the technique is simple

and uses less amount of memory during compression-decompression processes; thus making it suitable for mobile devices.

In DLWIC, the image is first converted into wavelet domain using orthonormal Daubichies wavelet transform [17] and the transformed data is then coded using QM coder, an advanced binary arithmetic coder [18]. The algorithm processes the bits of wavelet transformed image data in decreasing order of their significance in terms of MSE. This produces a progressive output stream; which allows the coding to be stopped when the quality of the reconstruction exceeds the threshold given as a parameter to the algorithm.

III. SEGMENTATION AND FEATURE EXTRACTION

Recently, with the rapid development of biomedical imaging technology, image segmentation technology has become the key technology of the medical image processing and analysis in order to extract the information from the medical images and to do the various following processing accurately. Some of the recent techniques for segmentation and extracting features of medical images have been discussed in this section.

A. Segmentation Algorithm Based on Multi-wavelet Analysis The multiresolution analysis of the multi-wavelet [19]-[21]

for the threshold value selection not only takes into account the impact of spatial distribution of images, but also the threshold transacting through the transformation coefficients which is on different resolution levels could restrain the noise, so it could improve the sensitivity of the segmentation results to the noise. Apply an orthogonal multi-wavelet to the medical image to do the threshold transacting and segmentation processing with the different wavelet coefficients on different resolution levels. Apply precise threshold value which is automatically determined by the multi-scale wavelet transform, to split the target organizations from the background. Better image feature information could be obtained by combining the auto segmentation method with the interactive segmentation method together [22]. B. Segmentation Based On Improved Genetic Algorithm (GA) Otsu algorithm [23] is one of the popular threshold algorithms, but because of the higher complexity of the algorithm itself, it is too time-consuming to apply for the real time image processing. Ting-lei Huang and Xue Bai [24] suggest a way of medical image segmentation using optimized two-dimensional Otsu method based on improved genetic algorithm (GA). In the proposed algorithm, the probability-ties of crossover are adaptively varied depending on the ranking value of individuals instead of fitness, and dyadic mutation operator was presented to take the place of the traditional one. The experimental results show that the new optimized method dramatically reduces the operating time in medical image segmentation while ensures the final image segmentation quality.


Page 2 of 348

C. Segmentation with Canny Operator and GVF Snake Model Canny edge detection algorithm [25] has been extensively

preferred for segmenting the images (particularly the human medical images that have high noise) because of its anti-noise ability. But, the edge based on Canny operator is not consecutive. Further, Xu and Prince have proposed a new deformable model called gradient vector flow (GVF) snake [26], [27]. Instead of using image gradients as the external force, it uses a spatial diffusion of the gradient of an edge map of image. However, But GVF still not can capture object contours in some complicated images. In [28] Jinyong Cheng et al. suggest a segmentation algorithm based on Canny operator and GVF snake model. According to the algorithm, first the rough edge is obtained by Canny operator. Then the thinning methods based on mathematical morphology have been used to get thin edge map of the image. With the edge map, GVF Snake can capture the accurate object boundaries. This method solves the problem that the edge based on Canny operator is not consecutive. And it improves GVF Snake model’s anti-noise ability. Thus this algorithm exploits the salient characteristics of both Canny operator and GVF model to improve the segmentation performance.

To carry on further Zheng Ying et al. in [29] have suggested a segmentation algorithm based on wavelet transform and improved generalized GVF (IGGVF) model. Firstly, wavelet transformation is carried out on the original medical image to get multi-scale reconstructed approximate images. Next a new initial setting method is employed for gaining the initial contour then it is deformed according to the IGGVF snake model to attain the ultimately rough contour in the largest reconstructed image. Afterwards, this contour is considered as the initial contour and continues to be deformed in smaller scale reconstructed image.

D. Segmentation Based on K-Means Clustering and Improved Watershed Algorithm The watershed segmentation technique has been widely

used in medical image segmentation e.g. to segment gray and white matter from MR images [30], [31]. Although the watershed transform is a fast, simple and more importantly, it is able to produce a complete division of the image in separated regions even if the contrast is poor, its drawbacks will include over-segmentation and sensitivity to noise. The fuzzy C-means clustering algorithm (FCM) is a soft-segmentation method that has been used extensively for segmentation of MR images [32]. However, its main disadvantages include its computational complexity and the fact that the performance degrades significantly with increased noise. K-means clustering algorithm [33], [34], on the other hand, is a simple clustering method with low computational complexity as compared to FCM. The clusters produced by K-means clustering do not overlap.

In [35] H.P. Ng et al. address the drawbacks of the conventional watershed algorithm when it is applied to medical

images by using K-means clustering to produce a primary segmentation of the image before applying the improved watershed segmentation algorithm [36] to it. The K-means clustering is an unsupervised learning algorithm, while the improved watershed segmentation algorithm makes use of automated thresholding on the gradient magnitude map and post-segmentation merging on the initial partitions to reduce the number of false edges and over-segmentation. The authors show that their proposed methodology produced segmentation maps which have 92% fewer partitions than the segmentation maps produced by the conventional watershed algorithm (see Fig. 3).

Tobias Klinder et al. in [37] have proposed a comprehensive solution for automatically detecting, identifying, and segmenting vertebrae in CT images as for many orthopaedic, neurological, and oncological applications, an exact segmentation of the vertebral column including an identification of each vertebra is essential.

IV. REGISTRATION, RETRIEVAL AND FUSION

Usually, physicians or radiologists examine medical images in conventional ways based on their individual experiences and knowledge. There is a big potential to automate this process and an ideal scenario is where a medical image (or image series) generated by an imagining station will be automatically (possibly in real time) compared with existing (reference or target) images stored in a database, then possible abnormalities may be identified and suggested by the system. With such capabilities, role of medical imaging would expand and the focus could shift from generation and acquisition to more effective post processing, organization, and interpretation. To approach this kind of automation two main technologies, namely image retrieval and image registration need to be addressed and integrated in a computer assisted diagnostic environment [38].

Image retrieval whether CBIR or text-based is the technique to find similar images from an image archive with the help of some key attributes associated with the images or features inherently contained in the images. Image retrieval is gaining importance as a support tool for diagnosis, research and education in medical domain [39]. Registration is another important technique in medical domain, generally needed for combining information from multiple imaging modalities, monitoring changes or evolution, image guided surgery or compare individual’s anatomies to standard atlas [40]-[42]. Generally, registration is referred to as the establishment of correspondence between images or between an image and physical space, which involves finding the transformation that brings different images of the same body part into strict spatial and temporal congruence [43].

Image fusion is the process of extracting meaningful visual information from two or more images and combining them to form a new image. In [44] authors propose the application of curvelet transform [45] in medical image fusion. In medical


Page 3 of 348

image fusion multiple medical images such as CT and MRI are fused into a new image to improve the information content for diagnosis. Some attempts have been proposed to fuse CT and MRI images using wavelet transform [46], but since medical images have several objects and curved structures it is expected that the curvelet transform would be better in their fusion.

V. TECHNOLOGIES EMBEDDED WITH MEDICAL IMAGING

• The design of software platform for medical imaging application has been increasingly prioritized as the sophisticated application of medical imaging. Jie Tian et al. [4] have designed and implemented a novel software platform in traditional object-oriented fashion with some common design patterns. This platform integrates the mainstream algorithms for medical image processing and analyzing within a consistent framework, including reconstruction, segmentation, registration, visualization, etc., and provides a powerful tool for both scientists and engineers.

• Traditionally, telemedicine services are provided by using wired networks such as telephones, and DSL or cable-modem-based broadband access systems to transmit biomedical data between a hospital and the point of care. However, these fixed systems have limitations in providing services to patients in remote localities and when the patients are mobile. Therefore, mobile telemedicine services with applications in emergency healthcare, telecardiology, teleradiology, telepathology, teledermatology, and tele-oncology have become popular to provide prompt and effective patient care. With emerging wireless technologies, patients can access healthcare services not only from hospitals, but also from rural healthcare centers, ambulances, ships, trains, airplanes, and homes. D. Niyato et al. [47] propose the application of IEEE 802.16/WIMAX-based broadband wireless for telemedicine/e-health services.

VI. CONCLUSIONS We have reviewed some of the recent advances and trends in

the area of biomedical imaging but it is certainly not complete. The accelerating pace of technological advances in the field of biological and medical imaging remains high, revealing more information about organs as well as function, place’s more demands on the performance of existing algorithms and new systems. Real-time multimodality imaging and image-guided interventions are the other fast growing areas of interdisciplinary research and development. Despite these wonderful advances, mainstream topics such as segmentation, registration and spatial-temporal integration still remain open and require breakthroughs to improve reliability and automation. Microelectromechanical systems (MEMS) employed by the medical device industry, appear to be destined for a flourishing biomedical imaging future.

Figure 1. (a) Original MRI brain image (b) ROI outlined, (c) compressed image enhanced with ROI coding [13]

Figure 2. Original and Resultant image of segmentation based on multi-wavelet analysis [22]

(a) (b)

(c) (d)

Figure 3. (a) Original MR image; (b) After K-means clustering; (c) Segmentation using traditional watershed algorithm (2756 partitions); (d) Final segmentation (172 partitions) [35]


Page 4 of 348

REFERENCES [1] A. Souza, J. K. Udupa, and P. K. Saha, “Volume rendering in the

presence of partial volume effects,” IEEE Trans.Med. Imag., vol. 24, no. 2, pp. 223– 235, Feb. 2005.

[2] A. Madabhushi and J. K. Udupa, “Interplay between intensity standardization and inhomogeneity correction in MR image processing,” IEEE Trans. Med. Imag., vol. 24, no. 5, pp. 561–576, May 2005.

[3] Y. Zhan and D. Shen, “Deformable segmentation of 3-D ultrasound prostate images using statistical texture matching method,” IEEE Trans. Med. Imag., vol. 25, no. 3, pp. 256–272, Mar. 2006.

[4] Jie Tian, Jian Xue, Yakang Dai, Jian Chen, and Jian Zheng, “A Novel Software Platform for Medical Image Processing and Analyzing”, IEEE Transactions on Information Technology in Biomedicine, vol. 12, no. 6, November 2008.

[5] Barna Iantovics, “The CMDS Medical Diagnosis System”, Ninth International Symposium on Symbolic and Numeric Algorithms for Scientific Computing, 2008.

[6] David Wu Et Al., “Perceptually Lossless Medical Image Coding”, IEEE Transactions On Medical Imaging, Vol. 25, No. 3, March 2006.

[7] D. Taubman and M.W. Marcellin, JPEG2000: Image Compression Fundamentals, Standard and Practice, 1st ed. Norwell, MA: Kluwer Academic, 2002.

[8] D. M. Tan, H. R. Wu, and Z. Yu, “Perceptual coding of digital monochrome images,” IEEE Signal Process. Lett., vol. 11, no. 2, pp. 239–242, Feb. 2004.

[9] D. M. Tan and H. R.Wu, Perceptual Image Coding, H. R.Wu and K. R. Rao, Eds. Boca Raton, FL: CRC, 2005, Digital Video Image Quality and Perceptual Coding.

[10] D. Wu, D. M. Tan, and H. R. Wu, “Visually lossless adaptive compression of medical images,” in Proc. 4th Int. Conf. Information, Communications & Signal Processing and 4th Pacific Rim Conf. Multimedia, 2003, pp. 458–463.

[11] D. Wu, D. M. Tan, and H. R. Wu,, “Vision model based approach to medical image compression,” in Proc. Int. Symp. Consumer Electronics, 2003, Paper ISCE3058.

[12] A. Said and W. A. Pearlman, “A new, fast and efficient image codec based on set partitioning in hierarchical trees,” IEEE Trans. Circuits Syst. Video Technol., vol. 6, no. 3, pp. 243–250, Jun. 1996.

[13] C. Doukas and I. Maglogiannis, “Region of Interest Coding Techniques for Medical Image Compression”, IEEE Engineering In Medicine And Biology Magazine, September/October 2007.

[14] M.A. Ansari and R.S. Anand, “Context Based Medical Image Compression with Application to Ultrasound Images”, IEEE India International Conferenec (INDICON), pp. 23-28, 2008.

[15] C. N. Doukas et al., “Medical Image Compression using Wavelet Transform on Mobile Devices with ROI Coding Support”, Proceedings of the 27th Annual IEEE Engineering in Medicine and Biology Conference, Shanghai, China, September 1-4, 2005.

[16] Lehtinen J., Limiting Distortion Of A Wavelet Image Codec, Acta Cybernetica 14 (1999), pp. 342-356.

[17] M. Vettereli and J. Kova~cevi’c, Wavelets and Subband Coding, Englewood cliffs, NJ, 1995.

[18] W. Pennebaker and J. Mitchell, JPEG: Still image data compression standard, Van Nostrand Reinhold, 1992.

[19] Claudio Rosito Jung, “Multiscale image segmentation using wavelets and watersheds”, IEEE, 2003.

[20] Mallatt S., and Huang WL, “Singularity detection and processing with wavelets [J]”, IEEE Trans. Information theory, 1992, 38: 617-643.

[21] Cotronei M, Montefusco LB, Puccio L., “Multiwavelet analysis and signal processing”, IEEE Transactions on Circuits and systems—II, 1998,45(8):970~987.

[22] Zhao Wencang and Wei Hongli, “Study of Medical Image Segmentation Algorithm Based Multi-wavelet Analysis”, Proceedings of the IEEE International Conference on Automation and Logistics Qingdao, China September 2008.

[23] Liu Jianzhuang, “A Two-dimensional OTSU Auto Image Segmentation Method for Gray-level Image”, Automation Journal,

Feb.1993, pp. 101-105. [24] Ting-lei Huang and Xue Bai, “An Improved Algorithm for

Medical Image Segmentation”, 2nd IEEE International Conference on Genetic and Evolutionary Computing, 2008.

[25] J. Canny. “A computational approach to edge detection,” IEEE Trans. on Pattern Analysis and Machine Intelli, Vol. 8, pp. 679-698, 1986.

[26] Xu, J. Prince, “Generalized gradient vector flow external forces for active contours,” Signal Processing. vol. 71, pp. 131–139, 1998.

[27] Xu, J. Prince, “Snakes, shapes, and gradient vector flow,” IEEE Trans. on Images Processing, vol. 7, pp. 359–369, 1998.

[28] Jinyong Cheng et al., “Segmentation of Medical Images with Canny Operator and GVF Snake Model”, 7th World Congress on Intelligent Control and Automation, June 25 - 27, 2008, Chongqing, China.

[29] ZHENG Ying, “Medical Image Segmentation Based On Wavelet Transformation And IGGVF”, IEEE 2008

[30] V. Grau, R. Kikinis, M. Alcaniz, S.K. Warfield, “Cortical gray matter segmentation using an improved watershed transform”,

Proceedings 25th Annual Int. Conf. of Engineering in Medicine and Biology Society, Vol.1, 2003, pp. 618-621

[31] V. Grau, A.U.J. Mewes, M. Alcaniz, R. Kikinis, S.K. Warfield, “Improved watershed transform for medical image segmentation using prior information”, IEEE Transactions on Medical Imaging, Vol.23 (4), 2004, pp. 447-458.

[32] J.C. Bezdek, L.O. Hall, L.P. Clarke, “Review of MR image segmentation techniques using pattern recognition”, Medical Physics, Vol. 20(4), 1993, pp. 1033-1048.

[33] J. B. MacQueen, “Some methods for classification and analysis of multivariate observations”, Proceedings 5th Berkeley Symposium on Mathematical Statistics and Probability , 1967, pp. 281-297.

[34] T. Kanungo, D.M. Mount, N.S. Netanyahu, C.D. Piatko, R. Silverman, A.Y. Wu, “An efficient k-means clustering algorithm: analysis and implementation”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 24 (7), 2002, pp 881-892.

[35] H.P. Ng et al., “Medical Image Segmentation Using K-Means Clustering And Improved Watershed Algorithm”, IEEE 2006

[36] H.P. Ng, S.H. Ong, K.W.C. Foong, W.L. Nowinski, “An improved watershed algorithm for medical image segmentation”, Proceedings 12th International Conference on Biomedical Engineering, 2005.

[37] Tobias Klinder et al., “Automated model-based vertebra detection, identification, and segmentation in CT images”, Elsevier Medical Image Analysis 13 (2009) 471–482.


Page 5 of 348

[38] Md. Mahmudur Rahman, Tongyuan Wang, Bipin C. Desai, “Medical Image Retrieval and Registration: Towards Computer Assisted Diagnostic Approach”, Proceedings of the IEEE IDEAS Workshop on Medical Information Systems: The Digital Hospital (IDEAS-DH’04).

[39] H. D. Tagare, C. Jafe, J. Duncan, “Medical image databases: A content-based retrieval approach”, Journal of the American Medical Informatics Association, 4 (3), 1997, pp. 184-198.

[40] David Hawkes, Registration Methodolgy – Introduction, Chap 2 of Medical Image Registration, edited by Joseph Hajnal, David Hawkes and Derek Hill, 2000

[41] Derek LG Hill,Philipp Batchelor; Registration Methodology: Concepts and Algorithms, Chap 3 of Medical Image Registration, edited by Joseph Hajnal, David Hawkes and Derek Hill, 2000

[42] Daniel Rueckert, Non-Rigid Registration: Concepts, Algorithms and Applications, Chap 13 of Medical Image Registration, edited by Joseph Hajnal, David Hawkes and Derek Hill, 2000

[43] Registration, edited by Joseph Hajnal, David Hawkes and Derek Hill, 2000

[44] F.E. Ali Et Al., “Fusion Of MR And CT Images Using Curvelet Transform”, 25th National Radio Science Conference, March 18-20, 2008, Egypt.

[45] M. Choi et al., “The Curvelet Transform For Image Fusion”, International Society For Photogrammetry And Remote Sensing Istanbul (ISPRS 2004), 35, Part B8, 59-64, 2004.

[46] J. N’u~nez et al., “Multiresolution-Based Image Fusion with Additive Wavelet Transform”, IEEE Transactions On Geoscience Remote Sensing, Vol. 37 No. 3, 1204-1211, May 1999.

[47] D. Niyato et al., “IEEE 802.16/Wimax-Based Broadband Wireless Access and Its Application for Telemedicine/E-Health Services”, IEEE Wireless Communications, February 2007.


Page 6 of 348

Abstract: Color based classification is widely used in food industry. Using this paper we propose a non destructive, efficient and novel technique of colorization and then propose an efficient quality analysis for the spices to be exported. Prior to export, Cuminum Cyminum L(Cumin) seeds are subjected to inspection for the purpose of quality control and grading. For quality grading, the cumin seeds are cleaned and screened manually hence the practice is tedious, time consuming and labor intensive. We have developed of a machine vision system for quality inspection using combined measurement analysis. The methodology involves collecting the RGB images and then converting them to binary images using histogram analysis. Some colorization procedures are involved for image enhancement to distinguish the species from the background. Then the combined measurements consisting of area, minor axis length are extracted from the ISEF applied colorized images. The quality evaluation of the specified samples is done on the basis of quality table so formed. Keywords: Quality control, Machine Vision, Cuminum cyminum L (Cumin Seeds), ISEF edge detection, combined measurements.

I. INTRODUCTION

In the world of automation intense research is in progress on application of electronic eye and nose in food, beverages and spice industry. Non destructive technique so used makes it possible to examine the quality of food without damaging it. In this area of food industry such type of quality evaluation is gaining more momentum. Image processing is one of the most commonly used for this purpose especially for the quality control of proteins, lipids and species etc. Indian spice industry accounts for 30% to 45% in world trade. Indian spice industry especially from Gujarat (Unjha) and Tamil Nadu provide quality species at competitive prices. To increase share in exports of the species the need of Indian manufacturers is to ensure consistency in supply, product, quality, pricing, and

Jalpa J Patel, Research Scholar,M.E.(Communication), is with the Electronics & Communication Engineering Department, G H Patel College of Engg & Technology, V V Nagar-388120, Gujarat, India (e-mail: [email protected]).

Chintan K Modi, Asst. Prof, is with Electronics & Communication Engineering Department, G H Patel College of Engg & Technology, V V Nagar-388120, and Gujarat, India. (e-mail: [email protected]).

Kavindra R. Jain, Lecturer, is with the Electronics & Communication Engineering Department, G H Patel College of Engg & Technology, V V Nagar-388120, Gujarat, India (corresponding author to provide phone: 09904467724; e-mail: [email protected].)

.

marketing producers are now incorporating latest methods and techniques to ensure higher quality of species and herbs. On the basis of such large new techniques the most user friendly and best approach is with the help of image processing. Quality control is essential in the food and spice industry and efficient quality assurance is becoming increasingly important [4]. Consumers expect the best quality at a competitive and affordable price, good shelf-life and high safety while spice inspections require good manufacturing practices. In this context, machine vision technologies might prove highly relevant, providing objective measurements of relevant visual attributes related to food quality and safety, such as, the shape or color or odor of a given good. Traditionally, in India quality inspection is performed by trained human inspectors, who approach the problem of quality assessment in two ways, observing by looking and feeling the whole bulk of cumin seeds. In addition to being costly, this method is highly variable and decisions are not always consistent between inspectors or from day to day [11]. This is, however, changing with the advent of electronic imaging systems and with the rapid decline in cost of computers [8], cameras, peripherals and other digital devices. In this type of environment, machine vision systems are ideally suited for routine inspection and quality assurance tasks. Backed by powerful state-of-the-art electronic technologies [17], machine vision provides a mechanism in which the human thinking process is simulated artificially. Kavindra Jain et al [22],[23], developed a quality evaluation method describing two major parameters for quantification by taking a RGB image directly. The quality value computed using the method proposed in [22] is not near to the true quality values. Up till now colorization was implemented in food by Liyanage C De Silva[24] but not in spice industry. So we have combined colorization with the method proposed in [22] which results in to quality value near to true quality values. In this paper the problem being faced by the spice industry, in particular cumin seeds exporters is discussed in Section 2. Section 3 discusses the materials and methods proposed for quantifying the quality of cumin seeds. Colorization is described in the same section. The proposed algorithm is also discussed in the same section. Section 4 discusses the quantification for the quality of cumin seeds Section 5 comprise of results and discussion. Section 6

Quality Evaluation of Cuminum Cyminum L (Cumin) Seeds Using Colorization

Jalpa J Patel 1 Chintan K Modi2 and Kavindra R Jain3 1,3Research Student (M.E. Communication) EC Department, G H Patel College of Engg. and

Technology, V V Nagar 2Asst. Prof. EC Department, G H Patel College of Engg. and Technology, V V Nagar


Page 7 of 348

concludes the paper.

II. PROBLEM DEFINITION Cuminum cyminum L (Cumin) is a small and slender annual herb, which grows up to a height ranging from 40-50 cm with many branches and linear, dark green leaves. The stem has 3-5 primary and 2-3 secondary branches. Flowers are very small, mostly pink and sometimes white in color. Seeds are small elongated, oval and attain light yellowish to greyish brown in color. The seeds are mostly used as condiments in the form of an essential ingredient in all mixed spices and in curry powder for favouring vegetables, pickles, soups, sausages, cheese and other preparations and also for seasoning of breads, cakes and biscuits. It is used in many Ayurvedic and veterinary medicines as carminative, stomachic, astringent and is useful against diarrhoea and dyspepsia. A Cuminum cyminum L (cumin) seed contains pedestals as shown in Figure 1. The cumin seeds having pedestals are of prime importance for quantifying quality. If these pedestals are incorrectly removed then it not only deteriorates the quality but it also increases its pungency. This may lead to microbial growth and blackening of kernel at that particular point. Microbial growth spoils the taste of food and is unfit for consumption. Other than pedestals the presence of foreign elements in cumin seeds also reduces its quality. This foreign element basically consists of top part of the seeds, stones and only sticks as shown in Figure 2. This problem leads us to think that if the quality of the Cuminum cyminum L (cumin) seeds can be quantified and graded automatically using machine vision system at the initial level itself then this product would become a part of export at a higher rate. In [22],[23], a quality evaluation method is proposed, describing two major parameters for quantification by taking a RGB image directly. The quality value computed using the method proposed in [22] is not

near to the true quality values.

III. MATERIALS AND METHODS In this section we discuss the proposed algorithms along with the definition of quality based on combined measurements technique. We have used area of cumin seeds and minor axis length of cumin seeds for counting the number of Cuminum cyminum L (cumin) seeds with long pedestals as well as foreign elements.

A. System Description, operating procedure and proposed algorithm: The system used is the same as described [21] which is shown in figure 3. The operating procedure is outlined in table I.

As per the procedure the person should select random samples from bulk of cumin seeds to be packed and sent to the market. These samples are spread on a tray in such a way that there is no overlapping of the seeds. From the image captured according to Table II quality of cumin seeds is determined and displayed on screen [22]. The simplicity of operation of system can be concluded from the operating procedure as in Table I.

Fig 3: Overall System

A 6M pixels camera with 3X optical zoom is used. To ensure proper luminance [4] for good quality of image, we placed the bulbs at point number 2 and 3 as shown in the Figure 3. Butter paper is used for uniform distribution of light on the tray. In one side of the box an opening is kept (point 5) to insert a tray containing cumin seeds for image capturing. We propose combination of colorization and the algorithm of [22] to evaluate the quality of cumin seeds. The proposed algorithm is outlines in Table II.

Fig 1: Cumin seeds with and without pedestals

Fig 2: Foreign Elements in the sample

TABLE I OPERATING PROCEDURE

Sr. No.

Steps

1 Spread the samples uniformly on the tray to avoid overlapping of seed.

2 Capture the image 3 Processing the image in computer 4 Display number of cumin seeds with pedestals

and foreign elements on screen. 5 Repeat the steps 1 to 4 for 10 to 15 samples


Page 8 of 348

An RGB image as shown in Figure 4 is converted in to gray image as shown in Figure 5. We apply colorization using swatching on the gray image then it is converted in to YUV. Foreground image is scribbled with one color while one seed out of all seeds is scribbled with anther color.

Fig 4: RGB image of the sample

Fig 5: Gray Scale image

Fig 6: Colorized image

Then swatching based colorization [3] is applied which results in to a scribbled image having all the seeds scribbled with same color. Optimization based colorization [3] is applied on this image and a final colorized image is obtained as shown in Figure 6. The following cost functions are optimized to get colorized image.

2

( )

( ) ( ) ( ) ...........(2.3.1)r s N s

J U U r WrsU s∈

= −

∑ ∑

2

( )( ) ( ) ( ) ...........(2.3.2)

r s N sJ V V r WrsV s

∈

= −

∑ ∑

2 2exp ( ( ) ( )) / 2 ...........(2.3.3)rWrs Y r Y s σ∝ − −

An optimal edge detection technique, ISEF, given

by Shen and Castan is applied on the colorized image [12] to separate the cumin seeds having long pedestal as well as foreign elements being present in the sample. Hysteresis thresholding is applied in the ISEF algorithm (Table III).

ISEF is an optimal edge detector like canny edge

detector which gives optimal filtered image. First the whole image will be filtered by the recursive ISEF filter in X direction and in Y direction, which can be implement by using equations given below. Recursion in x direction:

[ ] ( )( ) [ ] [ ],1,1,11,1 −+

+−= jiybjiI

bbjiy (1)

MiNj ..1,...1 ==

[ ] ( )( ) [ ] [ ]1,1,11,2 ++

+−= jiybjiI

bbbjiy (2)

MiNj ..1,1... ==

[ ] [ ] [ ]1,2,1, ++= jiyjiyjir (3)

Recursion in y direction:

[ ] ( )( ) [ ] [ ],,11,11,1 jiybjiI

bbjiy −+

+−= (4)

NjMi ..1,...1 ==

[ ] ( )( ) [ ] [ ],,11,11,2 jiybjiI

bbbjiy ++

+−= (5)

TABLE III ISEF ALGORITHM

No Steps 1 Apply ISEF Filter in X and Y direction 2 Apply Binary Laplacian Technique 3 Apply Non Maxima Suppression 4 Find the Gradient 5 Apply Hysteresis Thresholding 6 Thinning

TABLE II PROPOSED ALGORITHM TO COMPUTE QUALITY

Sr.No. Steps

1 Select the region of interest of the cumin seeds 2 Convert the RGB image to gray images

3 Apply swatch based colorization using histogram of gray level image.

4 Resultant is scribbled image 5 Apply colorization using optimization on scribbled

image to get a colorized image 6 Apply the ISEF edge detection on colorized image 7 Calculate the area of the cumin seeds.

8 Calculate the minor axis of the cumin seeds 8 Find the histogram of the areas of cumin seeds. 9 Compute the threshold value based on histogram 10 Classify cumin seeds based on the threshold value

computed from the histogram in to different grades. 11 Display the total number of cumin seeds in the

sample, number of cumin seeds with pedestals and number of foreign materials in the sample.


Page 9 of 348

NjMi ..1,1... ==

[ ] [ ] [ ]jiyjiyjiy ,12,1, ++= (6) b=Thinning Factor (0<b<1) The Laplacian image can be approximated by subtracting the filtered image from the original image. At the location of an edge pixel there will be zero crossing in the second derivative of the filtered image. The first derivative of the image function should have an extreme at the position corresponding to the edge in image and so the second derivative should be zero at the same position. And for thinning purpose apply non maxima suppression as it is used in canny for false zero crossing. The simple thresholding can have only one cutoff but Shen-Castan suggests to use Hysteresis thresholding. Spurious response to the single edge caused by noise usually creates a streaking problem. Streaking can be eliminated by thresholding with Hysteresis. Individual weak responses usually correspond to noise, but if these points are connected to any of the pixels with strong responses, they are more likely to be actual edge in the image. Such connected pixels are treated as edge pixels if there response is above a low threshold. Finally thinning is applied to make edge of single pixel. In Figure 7(a) cumin seed of good quality without pedestal is shown, while Figure 7(b) contains an image of a seed with pedestal. After applying the ISEF algorithm we get images of Figures 8 (a) and (b) respectively.

(a) (b)

Fig 7: Colorized Cumin seed with and without pedestals

(a) (b)

Fig 8: ISEF applied Cumin seed with and without pedestals

B . Computation of area and minor axis length of cumin seeds based on colorized image:

The area A of any object in an image is defined by the total number of pixels enclosed by the boundary of the object. While the minor axis length M of an image is defined as the length (in pixels) of the minor axis of the ellipse that has the same normalized second central moments as the region.

If we find the area of each cumin seeds separately and if the area of normal cumin seeds is, x, the area of cumin seeds with pedestal to be y will mostly be greater than x. While in case of foreign elements like sticks and

lower quality seed (thin seed), the area z will mostly be less than x. Use of minor axis length of the seed will remove the usage of Vernier caliper in case of human inspection system. To make the classification robust, automatic threshold computation based on histogram of area is used.

F. CLASSIFICATION OF CUMIN SEEDS: The classification of the cumin seeds is done based on combined measurements of area and minor axis. A typical histogram of area of cumin seeds computed from a typical sample is shown in Figure 9 to identify different clusters of cumin seeds. A typical histogram of minor axis length of cumin seeds computed from a typical sample is shown in Figure 10.

Fig 9: Histogram showing area of Cumin seed

Fig 10: Histogram showing minor axis length of Cumin seed

IV. QUALITY OF CUMIN SEEDS The classification of cumin seeds as cumin with pedestal, without pedestal and foreign elements is based on combination of two parameters that is the area and the minor axis length of cumin seed. As per the definition of [21] quality of cumin seeds using the following formula

1 2

,cQx x

=+

(7)

Where x1 and x2 are percentage analysis of various sample. Assuming c to be 100, Q Table can be prepared defining various Q values on the basis of two parameters x1 and x2 . The first row and first column of Figure12 represents percentage of x1 and x2 present in the bulk of Cuminum cyminum L (cumin seeds). In figure 11, the top most ellipse represents Grade ”A”, the red color filled area represents grade ”B” which comprise of higher percentage of foreign elements as well as cumin seeds with pedestals. The bottom most encircled area is grade ”C”.


Page 10 of 348

Fig 11: Q Curves according to figure 12

V. RESULTS AND DISCUSSION On the gray version of the colorized image, ISEF is applied in both x and y directions recursively. The filtered image is subtracted from the original image and the binary Laplacian image is obtained as shown in Figure 13.

Fig 13: Gradient image of the sample

Fig 14: Hysteresis image of the sample

After applying the positive and negative zero crossing on the gradient image it gets thinned. We apply Hysteresis thresholding with HT=30. The hysteresis threshold image is shown in Figure 14.

The thresholded image is analyzed for computation of areas of cumin seeds with and without long pedestals and foreign elements. The area of cumin seeds which are greater than 275 in terms of pixels are the Cuminum cyminum L (cumin seeds) with pedestals. Area of Cuminum cyminum L (cumin seeds) without pedestals is between 100 and 275 pixel square. The foreign elements being present are having areas less than 100 pixel squares

correspondingly. Correspondingly there are changes in minor axis length of Cuminum cyminum L (Cumin) seeds. The results so obtained by the analysis of eleven such samples are given in Table IV.

TABLE IV RESULT ANALYSIS OF VARIOUS SAMPLES

Sample Total seeds

Seeds with pedestal

foreign elements

normal seeds

1 53 10 2 41 2 48 8 3 42 3 51 19 0 29 4 52 12 2 37 5 49 11 3 35 6 51 16 6 29 7 53 14 2 37 8 56 10 3 43 9 51 11 3 38

10 48 5 1 42 11 55 4 3 48

Table V describes the percentage results of the

analysis of eleven such samples. We compare the results obtained by method of [22] being x1k, x2k and results obtained by proposed colorization based method being x1j, x2j , along with true values x1, x2. The percentage average of x1 is approximately 25.7 and that of x2 is 4.26, giving the Q value as 3.334335 which shows that the sample under evaluation is not the best quality cumin seeds. The percentage average of x1k is approximately 22.8 and that of x2k is 7.4, giving the Q value as 3.269 which also concludes that the sample under evaluation is of ‘B’ grade. But the method gives Q value which follows the Q5 line, i.e. bottom range of ‘B’ grade. The percentage average of x1j is approximately 17.79 and that of x2j is 6.98, giving the Q value as 4.03, which also shows that the sample under evaluation is not the best quality cumin seeds. A careful observation of mean absolute errors given in Table VI motivates to compute the Q value using x1k and x2j . Using the same we get Q value equal to 3.3545 which is very near to the true Q value of 3.334335.

VI. CONCLUSION

The method proposed in the paper gives a direction of certification of quality of cumin seeds based on non destructive machine vision based technique. It is better to compute number of foreign elements using colorization technique proposed in this paper to get accurate quality evaluation.

ACKNOWLEDGEMENT The authors are thankful to Prof. R. K. Jain for his valuable suggestions. We are also thankful to the Unjha market for providing us the samples of cumin seeds. We express our gratitude to Mr. Akul K Modi for making the entire system model.


Page 11 of 348

REFRENCES [1] Abdullah MZ, Fathinul-Syahir AS, Mohd-Azemi BMN

(2005)”Automated inspection system for color and shape grading of starfruit (Averrhoa carambola L.) using machine vision sensor.” Transactions of the Institute of Measurement and Control, 27 (2), 65-87

[2] Abutaleb AS (1989)”Automatic thresholding of grey-level pictures using two-dimensional entropies.” Pattern Recognition, 47 (1), 22-32.

[3] A. Levin, D. Lischinski, and Y. Weiss, “Colorization using optimization,” ACM SIGGRAPH 2004 Conference Proceedings, 2004.

[4] Batchelor BG (1985) ”Lighting and viewing techniques In AutomatedVisual Inspection”(Batchelor BG, Hill DA, Hodgson DC, eds). Bedford:IFS Publication Ltd, pp.103-179.

[5] Blasco J, Aleixos N, Molt E (2003) ”Machine vision system for automatic quality grading of fruit.” Biosystems Engineering, 85 (4), 415-423.

[6] Canny J (1986) ”A computational approach to edge detection.” IEEETransactions on Pattern Analysis and Machine Intelligence, 8 (6), 679-698.

[7] J. M.Aguilera, A.Cipriano, M.Erana, I.Lillo ,D.Mery A.Soto.”ComputerVision for Quality Control in Latin American Food Industry”

[8] Du C-J, Sun D-W (2006c) ”Learning techniques used in computer visionfor food quality evaluation: a review.” Journal of Food Engineering, 72(1), 39-55.

[9] Gan TH, Pallav P, Hutchins DA (2005) ”Non-contact ultrasonic quality measurements of food products.” Journal of Food Engineering, 77 (2),239-247.

[10] Ginesu G, Guisto DG, Mrgner V, Meinlschmidt P (2004)”Detection of foreign bodies in food by thermal image processing.” IEEE Transactions of Industrial Electronics,51 (2), 480-490.

[11] Harmond J E,Kelin L M and Brandenburg N R.(1961).”Seed Cleaner and handling”.USDA Handbook,179:38-48.

[12] R.P.Kachru et.Al.,”Development and testing of pedal cum power operated air screen cleaner.”Agrcultural Machanization in Asia,Africa and Latin America.21(4):29-32.1990

[13] Kapur JN, Saho PK, Wong AKC (1985)” A new method for gray level picture thresholding using the entropy of the histogram.” ComputerVision, Graphics, and Image Processing, 29, 273-285.

[14] Fu X, Milroy GE, Dutt M, Bentham AC, Hancock BC, Elliot JA (2005)”Quantitative analysis of packed and compacted granular systems by x-ray microtomography.” SPIE Proceedings Medical Imaging and ImageProcessing, 5747, 1955.

[15] Pun T (1980)”A new method for gray-level picture thresholding using the entropy of the histogram.” Signal Processing, 2 (3), 223-237.

[16] Paulsen M (1990)”Using machine vision to inspect oilseeds.”

INFORM,1 (1), 50-55.

[17] Pearson T (1996) ”Machine vision system for automated detection of stained pistachio nuts.” Lebensmittel Wissenschaft and Technologie, 29 (3), 203-209.

[18] Xiaopei Hu, ParmeshwaraK.M.,DavidV.”Development Of Non Destructive Methods To Evaluate Oyster Quality By Electronic Nose Technology”. Springer Science Business Media, LLC2008

[19] Du CJ, Sun D-W (2004a) ”Recent development in the applications of image processing techniques for food quality evaluation.” Trends in Food Science and Technology, 15, 230-249.

[20] Shen Castan, Sian Zhao,”A Comparitive study of Performance of Noisy roof edge detection ”5th International conference on Computer analysis of Images and Patterns, volu.179,pp 170-174

[21] Vishwanathan R.et.Al..,”Development of hand operated rotary sieve cleaner -cum- grader for seasum seed.”A.M.A,25(2):35-36,1994

[22] Kavindra R.Jain,Chintan K. Modi, Kunal J.Pithadiya, ”Quality evaluation in spice industry with specific refrence to cumin seeds “MONAISH UNIVERSITY CITISIA 2009(IEEE), Malyasia conference in JULY 2009.

[23] Kavindra R.Jain,Chintan K. Modi, Kunal J.Pithadiya, ”Non destructive quality evaluation in spice industry with specific refrence to foeniculum vulgare(fennel seeds)”, 13th International Conference,IASTED VIIP,2009, C ambridge,UK conference in July 2009.

[24] Liyanage C De Silva, Anton Pereira, and Amal Punchihewa Institute of Information Sciences and Technology, Massey University, New Zealand.

[25] T. Horiuchi and S. Hirano, “Colorization algorithm for grayscale image by propagating seed pixels,” Proceedings of IEEE International Conference on Pattern Recognition, 2003, pp. 457-60.

[26] Tomihisa Welsh, Michael Ashikhmin ,Klaus Mueller “Transferring Color to Grayscale Images” by Center for Visual Computing, Computer Science Department, at Stony Brook.

[27] Vishwanathan R.et.Al..,”Development of hand operated rotary sieve cleaner -cum- grader for seasum seed.”A.M.A,25(2):35-36,1994.

TABLE V COMPARISION OF PERCENTAGE WISE ANALYSIS

Sample

Total seeds

x1 (%)

x2 (%)

x1k (%)

x2k (%)

x1j (%)

x2j (%)

|x1-x1k| |x2-x2k|

|x1-x1j|

|x2-x2j|

1 53 15.09 3.77 20.75 3.77 24.52 3.77 14.24 0 19.33 0 2 53 15.09 1.88 13.20 7.54 9.43 3.77 7.30 7.30 11.78 3.26 3 49 38.77 0 38.78 0 22.44 6.12 0 0 31.61 6.12 4 52 25 0 11.54 13.46 13.46 3.84 22.17 13.46 21.06 3.84 5 48 29.16 6.25 31.25 6.25 16.66 6.25 11.21 0 23.93 0 6 51 29. 13.72 25.41 11.76 17.64 11.76 14.67 7.06 23.52 7.06 7 55 30.90 1.81 12.72 9.09 10.90 20 28.16 8.90 28.91 19.91 8 51 35.29 5.88 49.01 5.88 49.01 7.84 34.01 0 34.01 5.18 9 48 37.5 6.25 37.5 4.16 10.41 2.08 0 4.65 36.02 5.89 10 51 19.60 1.96 3.92 19.6 17.64 7.84 19.21 19.50 8.54 7.59 11 56 7.14 5.35 7.14 0.12 3.57 3.57 0 3.99 6.18 3.99

Average 25.72 4.26 22.83 7.42 17.79 6.98 13.72 5.90 22.26 5.71


Page 12 of 348

Proceedings of the 2009 International Conference on Signals ...

Documents

Transcript of Proceedings of the 2009 International Conference on Signals ...