2009 10 jan RFID 2

Transcription

2009 10 jan RFID 2
SEVENTH ANNUAL NATIONAL CONFERENCE
on
“BIOMETRICS,
RFID AND EMERGING TECHNOLOGIES
FOR
AUTOMATIC IDENTIFICATION”
th
DATE : 10 JANUARY, 2009
VENUE
Thorale Bajirao Peshwe Sabhagruha,
“Jnanadweepa” Thane College Campus, Chendani Bunder Road, Thane - 400601
Editorial Committee
Mrs. Suhasini Shukla
Mrs. Anita S. Diwakar
Mrs. Archana Mohite
Mrs. Reena Khairnar
Mrs. Neelima Gajare
Mrs. Medha Patki
Mrs. Vibhavari Ikhe
Mrs. Nandini Pusalkar
Ms. Swati Ganar
Supporting Staff
Mr. R. D. Patil
Mr. V. K. Vichare
Organised by
VPM’s POLYTECHNIC, THANE
‘Jnanadweepa’ Bunder Road, Chendani, College Campus,Thane (w) - 400 601
Tel. : 2536 4494, 2533 9872. URL - www.vpmthane.org
VISION : ENSURING QUALITY TECHNICAL EDUCATION
V. P. M’s POLYTECHNIC, THANE.
We take immense pleasure in cordially inviting you for the
Seventh National Conference
on
Biometrics,
RFID and Emerging Technologies for
Automatic Identification
Inaugurl Function
on Saturday 10th January, 2009 at 9-00 am
Chief Guest
Prof. G. B. Dhanokar
Director, Maharashtra State Board of Technical Education
Guest of Honour
Mr. T. Vasu
Director, Tandan Group of Companies
Keynote Speaker
Mr. Harish Aiyer
Director. Softcell Ltd.
Dr. V. V. Bedekar
Chairman, Vidya Prasarak Mandal, Thane.
will preside over the function
Prof. D. K. Nayak
Convener/Principal
V.P.M’s Polytechnic, Thane
Mrs. A. S. Diwakar
Jt. Secretary
Computer. Engg. Dept.
Mrs. S. K. Shukla
Organising Secretary
HOD - Computer. Engg. Dept.
Conference Venue
Thorale Bajirao Peshwe Sabhagruha
‘Jnanadweepa’, College Campus, Chendani Bunder Road, Thane (W) - 400 601
Maharashtra State, India.
Patron
Dr. V. V. Bedekar
Chairman,
Vidya Prasarak Mandal, Thane.
ADVISORY COMMITTEE
Dr. Siddhan
Mr. N. S. Barse
Business Advisor
Pharmalabs Limited
Mumbai
Librarian
K. G. Joshi College of Arts
Convener
Mr. N. K. Marwah
Prof. D.K. Nayak
Managing Director
Magna Services Pvt. Ltd
Principal,
VPM's Polytechnic, Thane
Mrs. S. K. Shukla
Mrs. A. S. Diwakar
Organising Secretary
HOD, Computer Engineering,
VPM's Polytechnic, Thane
Joint Secretary
Computer Engineering,
VPM's Polytechnic, Thane
Mr. Kaustubh Kale
Mr. Ashish Hastak
Anant Corporation
Manager, Airtel Ltd
iii
LIST OF ANNUAL NATIONAL CONFERENCES
ORGANIZED BY THE POLYTECHNIC IN THE PAST
Sr.
Name of the Conference
Date /Year
1
Pollution of Water Bodies in the Urban Areas
8th August 2004
2
Alternative Energy Sources
27th and 28th August 2005
3
Geo Informatics
8th and 9th December 2006
4
Innovations in Safety, Health & Environment
3rd February 2007
5
Latest Trends in Nanotechnology
5th January 2008
6
Corrosion Prevention Through
18th October 2008
No.
Advanced Technologies
Please note :
The Authors of the papers are alone responsible for technical contents of the papers and references cited
therein.
Published by :
V. P. M.’ Polytechnic,
“Jnanadweepa” Chendani Bunder Road,
Thane (W), Maharashtra
Tel. : 2536 4494
E-mail : vpm_polytechnic@rediffmail.com
URL : www.vpmthane.org
Printed by :
Jayesh Patel
“Three Art”
Hilari Compound, Left to Rajashree Tower,
Nr. Pratap Talkies, Kolbad, Thane (W) - 400601
Mo. : 98928 30581
Email : designer jayesh_67@rediffmail.com
iv
Message
Dear friends,
I have a great pleasure in presenting the proceedings of one day national conference
on “Biometrics, RFID and Emerging Technologies for Automatic
Identification”. It is the seventh national conference being organized by the
polytechnic with the aim of creating awareness on futuristic subjects amongst students
and teachers. The theme of the conference is the latest development in the field of
science and technology and is extremely relevant at this point in time when limitations of
manual tracking and locating a specific item speedily and without error are of prime
importance.
The conference theme is one of the most developing areas in Indian industries as they are increasingly
using these techniques in transportation, storage and manufacture in many areas like electronics, airlines,
malls, cargo services and plenty of new applications are coming up.
Automatic Identification is an interdisciplinary area with faculty from Mathematics, Physics and
Computer group playing their role. For students of this generation knowledge of Automatic Identification
techniques will definitely bring success.Emergence of Automatic Identification techniques for humans leads
to greater sense of security and hence is gaining more importance. Some of the identification techniques
currently being used in the industry are smart cards for attendance, RFID for supply chain management, RFID
for library book management.
The development of cost effective and reliable devices, their processes and their installation in newer
applications and fields will be the next step in the growth of Indian industry.
The seminar on “Biometrics, RFID and Emerging Technologies for Automatic
Identification” is our small step in imparting knowledge regarding this technology to one and all. I am
sure the dedicated and coordinated efforts of all of us will succeed in fulfilling the very objective of the
conference. I am sure the souvenir of the conference would help the participants in coming up with their own
innovative ideas and develop the identification techniques further.
With best wishes
Dr. Vijay V. Bedekar
Chairman
Vidya Prasarak Mandal, Thane.
v
National Conference on
Biometrics, RFID and Emerging Technologies for Automatic
Identification.
Message from the Director, MSBTE
I am delighted to know that V.P.M.'s Polytechnic has organized 7th National conference on Biometrics,
RFID and Emerging Technologies for Automatic Identification on 10th January 2009.
As we all know Automatic Identification offers a powerful platform that leads to various aspects such as
supply chain management using RFID, biometrics, smart cards etc. This presents immense opportunities to
various industries by way of improving the object monitoring system.
Realizing the importance and potential of Automatic Identification many industries are actively
implementing these technologies. RFID (Radio Frequency Identification) is one such method being used all
over the world in various applications such as vehicular tracking, healthcare. postal services etc. Even
educational institutes are using automatic identification techniques in various projects.
Automatic Identification techniques will gain further importance in the current scenario of global
recession. The techniques used for identification of human being by identifying characteristics such as finger
image, palm image, facial image, iris print etc will be fast gaining importance as the security issues come
forward.
MSBTE has always extended its support to such kind of national conferences. It organizes various
training programs for faculties for imparting current knowledge, technology and for curriculum content
updating.
I sincerely appreciate the efforts of organizers and hope that the publication of this souvenir will go long
way in fulfilling the very objective. I wish the conference all the success.
Prof. G. B. Dhanokar
Director,
Maharashtra State Board of Technical Education.
vi
25.12.2008
Preface
Dear Participants,
Kindly accept my Season's Greetings. The Seventh National Conference
Organized by Polytechnic is focusing on Emerging trends in Automatic
Identification. The IT revolution has created extensive dependence on technology
infrastructure for every Business Enterprise and it has become the backbone of the
industry. With the Business activities getting integrated into Latest technologies,
Advanced security features are finding their place for reliability, Integration, and
Productivity enhancement.
Today Information Security cannot be ensured just with traditional methods. In addition to conventional
methods, features are being introduced to make Authentication accurate. Biometrics has become essential
part of HR Management in Organizations to maintain accountability of staff services. RFID is used in
Resource & Materials Management in Stores, Libraries, Airlines, Cargo systems and many Global
applications. These technologies help to reduce inaccuracies in the Information Processing.
The Conference theme is chosen to enrich the students especially from Electronics and Computer group
Diploma Programmes. These are the sunrise technologies in ITES and will be used in large scale in
Organizations. New technologies of identification will help the students and professionals to continue their
innovative instincts for adding features to existing technologies.
The invited speakers will definitely be enthused to interact with the participants to share their knowledge
& expertise. It will be an enriching experience at the end of the Conference to visualize the emerging
technology with renewed perspective.
All the invited speakers, & exhibitors contributions are note worthy and they deserve compliments. On
behalf of the Polytechnic. I convey my sincere gratitude's for their valuable support to this conference,
Special thanks to Prof. P. A. Naik Secretary, Prof. G. B. Dhanokar, Director Maharastra State Board of
Technical Education for their Valuable support and sponsorship.
Sincere thanks to Dr. V. V. Bedekar, Chairman & all members of V.P.M for their encouragement in
organizing such National Conferences.
Warm Regards,
Prof. D. K. Nayak
Principal & Convener
vii
Biometrics, RFID and Emerging Technologies for
Automatic Identification
It gives me great pleasure to present this souvenir to you that covers various issues
ranging from RFID and its applications, Standards developed for RFID use,
Biometrics applications such as face detection, signature recognition to Fuzzy
decision trees etc.
Automatic Identification refers to the methods of automatically identifying
objects, collecting data about them and entering the data into computer system without
human intervention. Technologies considered to be part of Automatic Identification
are RFID, Biometrics, and Smart Cards etc.
Automatic Identification has become the need of the day for various reasons. The first and the foremost
reason is the global recession that the world is now facing. Secondly the security aspect of data and the
authentication of user become essential in the current scenario.
The most commonly used automatic Identification techniques being RFID (Radio Frequency
Identification) which is being used in various applications such as Automatic Vehicular Tracking, Library
Management, and Supply Chain Management etc.
Another common technique used in automatic identification is the use of smart cards which are being
introduced by almost all the companies for staff attendance.
Biometrics with finger print, iris, palm, voice, signature recognition has become essential when we look
at the current situation in the world.
Necessity is the mother of all inventions: Greater demands for new products is so ever increasing that it is
necessary for the industry to keep up with the requirements. However we must also keep in mind the negative
aspects of Automatic Identification such as initial high investment, dependence on other technologies such as
sensor readers for security, privacy concerns etc.
As an organizing secretary I express my sincere thanks to all the speakers for accepting our invitation,
delegates for their response and the sponsors for their generosity. I am also grateful to the authors who have
contributed to the souvenir to augment its utility for the students, technical teachers and practitioners in the
field.
I wish to express my gratitude Mr. Marwah for his help and encouragement.
I am thankful to our Chairman Dr. V.V. Bedekar and Principal, Prof D.K.Nayak, who provided the needed
academic and other support and were constant source of encouragement to me in facing the challenges in the
organization of the seminar. Thanks are also due to my colleagues for their constant support and
encouragement.
Mrs. Suhasini Shukla
Organizing Secretary
H.O.D.
Department of Computer Engineering
viii
V. P. M's Polytechnic Thane.
7th ANNUAL NATIONAL CONFERENCE
On
BIOMETRICS, RFID AND EMERGING TECHNOLOGIES FOR
AUTOMATIC IDENTIFICATION
(10th January 2009)
Program Schedule
Session
Name of Person
Inaugural
Registration
Session
Inaugural Function
Chief Guest :
Prof. G. B. Dhanokar
Director, MSBTE.
Guest of Honour :
Mr. T. Vasu, Tandan Group of
Companies
Dr. V. V. Bedekar, Chairman,
Vidya Prasarak Mandal, will
preside over the function.
9.00 - 10.00am
Keynote Address
RFID
Mr. S. Kalyanaraman,
G.M. (Engg.) Tandan Group of
Companies
Biometrics :
Mr. Harish Aiyer,
Director, Softcell Ltd.
10.00 - 10.45 am
TEA BREAK
Plenary Session Presentations by dignitaries
from Industry
Chair Person :
Dr. Siddhan
Session II
Chairperson
Mr. Sameer
Phanse
LUNCH BREAK
Paper 1
Paper 2
Paper Details
Time
08.00 - 10.00 am
Industry Perspective of
Biometrics :
M. Anand, GM, Godrej & Boyce
RFID :
Mr. Nalawade, IIT Mumbai.
RFID : A tool for Library Security
Mr. N. S. Barse, Librarian.
K. G. Joshi College of Arts.
Multimodel Recognition based on
Fuzzy Decision Trees
Susmita Deb and M. Gopal
Dept. of Electrical Engg. Indian
Institute of Technology, Delhi.
ix
10.45 - 11.00 am
11.00 - 1.00 pm
01.00 - 2.00 pm
02.00 - 2.20 pm
02.20 - 2.40 pm
Paper 3
Session III
TEA BREAK
Technical Presentation
Paper 4
Chairperson
Prof. Uday
Pandit Khot
Paper 5
Paper 6
Valedictory Function
Specialized GlobalFeatures for
Off-line Signature Recognition
H. B. Kekre, V. A. Bharadi,
Dept. Of Computer Engg., Dept.
Of E & TC., Thadomal Shahani
Engg. College.
2.40 - 3.00 pm
3.00 - 3.20 pm
3.20 - 3.40 pm
3.40 - 4.00 pm
Gunnebo Industries
Fingerprint Recognition in DCT
Domain,
M. P. Dale, S. N. Dharwadkar and
M. A. Joshi (HOD) COEPMES’s
College of Engg., Pune.
Automatic Identification
4.00 - 4.20 pm
Techniques Frequency Ranges
and Radio Liscensing Regulations
Rahul Kulkarni, Siemens.
Kirti Agashe, HOD, IE Dept.
V.P.M’s Polytechnic, Thane
RFID, Its Applications and
4-20 - 4.40 pm
Integrating RFID in WSN
Anuradha Kamble and Anita S.
Diwakar, Lecturer, SIES Institute
Of Technilogy, Nerul, NaviMumbai, V.P.M.’s Polytechnic.
Thane.
Chief.Guest :
4-40 - 5-00 pm
Mr Marwah MD, Magna Services.
x
Mr. T. Vasu
Profile
A Mechanical Engineering graduate from the University of Madras.
Mr. T. Vasu, started his professional career with a public sector company in
Madras making teleprinters and other telecom related equipments, from late
60's.
Having joined in the field of Electronic Manufacturing, he got trained in
Electronics with M/s. Olivetti, Ivrea, Italy, for 6 months and was awarded a
technical diploma. Had exposure in various functions of the business
operations like manufacturing, quality, supply chain, customer service,
Industrial Engineering & Systems taking part in the management discussions
as well moving up in the higher echelons in Management before he joined
Tandon Group, a large Hardware manufacturer- exporter at their Madras Export Processing Zone (MEPZ)
plant in 1989, as the Head of Operations.
While in MEPZ, he contributed to the exporting community by serving as Honorable Secretary to the MEPZ
Manufacturer's Association.
During the period, he was also a member of the Gill Committee formulated by the Ministry of Commerce to
go into the simplification of various procedures related to the Export Processing Zone in the country.
After serving the Celetronix Group Companies in various capacities for including Managing Director of
Celetronix Power India Pvt. Ltd., and jabil Circuits for nearly 20 years, currently serving
Tandon Group in the capacity of Director, Corporate Affairs.
Mr. Vasu is the immediate past Chairman,
Export Promotion Council for Export Oriented Units and Special Economic Zones (EPCES) sponsored
by the Ministry of Commerce of Industry, Government of India, and
Vice President, Elcina Electronic Industries Association.
xi
S. Kalyanaraman
General Manager (Engineering)
Tandon Group of Companies,
SEEPZ-Special Special Economic Zone, Mumbai.
Profile
Starting the career as Engineer (design and development) in Modular
electronics Chennai (India) which made head gimble assemblies,
headstacks, voice coils and hard disc drives in high voloume.
Moved up in the value chain as Senior Engineer and currently functions as
Gen. Manager (Engineering) in design and development services in the areas
of
1.
2.
3.
4.
5.
6.
7.
8.
Combination of Solar- Stirling Engine based Power Generation.
Concentrated Solar Photo voltaic technology.
Various Defense Products.
Test Equipments for Mobile Tower Antenna Testing.
ATE for UPS & Panel accessories Testing.
Compressed Air motor.
Tyre Pressure Management system for Trucks & Cars.
Low cost method for Silicon extraction for Photo voltaic.
Key expertise
Overall 18 years of hands on experience in the field of Design, Automation, Integration, Maintenance in
Electronics and Communication.
 Design & Development of Products for High End Car Audio Digital Amplfiers,
 Development of RFID Active & Passive Tags, Set top box related..
 Design and development of various Test equipments in high volume manufacturing of Hard disc
drive &
 Multiple mode Switchmode Power Supplies and Set top Boxes..
 Data acquisition systems.
 Combi tester. ELCINA AWARD WINNING PROJECT.
 Customer Level Tester.
 Fully automated and temp./humidity controlled Burn in systems.
 Power supply and front panel testers.
 Static tester for Headstack assembly.
 Quasi static tester for Headstack assembly.
 PLC Based swaging machine.
 PLC based Pivot Installation machine.
 Alarm Annunciate for ESD Outbreak.
 SCADA system for facilities department.
 Personal Monitoring system for Cleanrooms.
 ESD Monitoring system.
 DAS system
 Static R/W Tester.
 Static Attitude Tester.
xii
Harish Aiyer
Email : haiyer@gmail.com Mobile 9821013093
Residence: Plot 100, Sector 8 Vashi Navi Mumbai 400 703
Career Steps :
u Director, Softcell Technologies, Ltd. (2000-Present);
u President, Softcell T & T Inc USA (1998-2000);
u Executive Director, Somatico (1988-1998),
u Tech Support Manager, Sonata Software, Systems analyst, Team
leader, Project leader Unisys (1985-1988) Holland/ France/ USA/
Canada.
Business Background :
Softcell : (current) One of founders of Softcell Technologies, Ltd.
in 1989, and currently one the board of Director. Turnover of over
350 crores 2007-08and expected to cross Re 350 cores 2006 07.
Specializes in the BFSI systems and an expert in Credit Bureau
Operations.
Somatico Laboratories Pvt Ltd : (Pharmaceutical company) till 1999.
Harish as Executive Director was instrumental in setting up the entire manufacturing Somatico laboratories
PVT ltd. and Somatico Pharma Pvt ltd. As the executive director of Somatico
Laboratories Pvt Ltd
Other Professional Exposure :
 He has been a visiting / guest lecturer in many colleges and universities on
Corporate Strategy, Computer Science, Marketing, Knowledge Management,
Brain yoga and Change management. Personal development.
 He has extensive knowledge on Software architecture, Expert systems, ERP
Solutions, Health care systems,
Financial and Banking applications, Group ware systems and knowledge
Management.
 He is closely connected with corporate governance and Strategy formulation
with at least half a dozen of ltd
companies as part of their Board.
 Working with over a dozen international startup organizations as Advisor.
Associations & Honors :
 Bharatiya Vidya Bhavan's: Hon. Secretary Navi Mumbai Kendra,
 National Center for Leadership founder.
 Board of Management of SIES college of Management,
 Brain Yoga Institute Managing Trustee,
 Joint Secretary Awaz, An NGO helping Citizens issues of Navi Mumbai.
 Member of International Association of Business Leaders NC USA since 2001,
 Member of international who's who historical society London since 2003.
 Ex Vice Chairman of Bombay Management Association Navi Mumbai chapter,
 Mahatma Gandhi Institute of Computer Education, founding Committee Member;
 Ex president of all India association of college going scientist.
 Ex Sec of IES education Society's Alumni Association.
xiii
Education :
SP Jain Institute of Management, Master's Degree in Management Science (1985); APICS, Certification
(USA); The Project Management Institute, Certification (USA), internal auditor for CMM 1999, Lead
auditor for ISO 9000, Knowledge Management from Harvard 1999-2000. Certificate in Intellectual Property
laws 2006.
Personal Information :
Mr. Aiyer enjoys reading in his personal time. He practices martial arts, meditation and Brahma Vidya. He is
an amateur Astronomer, archeologist and Photography.
He is currently involved in research on Brain, Brain Development and Yoga. He has interests in Robotics,
Artificial intelligence and Knowledge management. Owns a personal library of over 7500 books [non fiction]
on various subjects.
Few of the invites for Lectures :
1.Speaker at the Conference on the issues on “year 2000” and computer systems in 1999 in Chicago USA.
2.Chief guest for the IEEE conference of Wireless technology in India 2007.
3.Speaker at the international seminar on Project management at KL Malaysia 2006
4.Key note speaker for National Seminar for Security NSA, IEEE 2008,
5.Chief Guest for launching the Management Courses by MET Mumbai 2008.
6.Chief Guest For launching the Management courses in SIES COMS 2008.
7.Speaker at the international Entrepreneurial forum in Singapore University 2007.
Mr. Marwah
Profile
A passionate recruitment Professional with over 20 years of extensive
Executive Search and Career Counseling experience, Mr. Marwah is Founder
& Managing Director of Magna Services (India) Pvt.Ltd., a company engaged
in providing ethical & innovating executive search and Recruitment consulting
in India and overseas..
Before Founding Magna, Mr.Marwah was the Dy CEO & Vice President of
Datamatics Staffing Services.Mr. Marwah, launched world's first job portal
like e services Employers Club in 1988 he also ran an Employment Newspaper
Employers' Dispatch & Courier. Mr. Marwah also writes on Recruitment
related subjects and his articles have appeared in some of the leading
newspapers.
During his two decades of profession, Mr.Marwah has worked in Asia, Africa and USA and was also
associated with organizations like Magna industries Australia, SM Dyechem, Empire Industries and NES UK
in different capacities.
Currently He is also engaged in incubating an innovating e- Human Capital Exchange which can
revolutionize recruitment by integrating and automating many of the recruitment processes
Mr Marwah is also an active Rotarian and member of NHRD and ERA other associations
Mr. Marwah is the Head Special Interest Group & member of Executive Council of Executive Recruiters
Association.
xiv
Contents
1
Industry perspective of Biometrics
M.Anand, GM, Godrej & Boyce
1
2
RFID: A tool for Library Security
Mr. N Barse, Librarian, K.G.Joshi College of Arts
3
3
Multimodel Recognition based on Fuzzy Decision Trees
Susmita Deb and M. Gopal,
Department of Electrical Engineering, Indian Institute of Technology, Delhi
6
4
Specialized Global Features for Off-line Signature Recognition
H B Kekre,V A Bharadi, Dept of Computer Eng, Department of E & TC.
Thadomal Shahani Engineering College
10
5
Fingerprint Recognition in DCT Domain
M.P.Dale, S.N.Dharwadkar and M.A.Joshi (HOD) COEP ,
MES's College of Engineering, Pune
16
6
Automatic Identification Techniques Frequency Ranges and
Radio Liscensing Regulations
Rahul Kulkarni, Siemens.
Kirti Agashe, HOD,Department of IE,V.P.M's Polytechnic,Thane
20
7
RFID, Its Applications and Integrating RFID in WSN
Anuradha Kamble and Anita S Diwakar,
Lecturer, SIES Institute Of Technology, Nerul, Navi Mumbai,
V.P.M's Polytechnic,Thane
25
8
A Skin Color Segmentation Approach towards Face Detection
Ashwani Kumar,
Dept. Of Electrical and Instrumentation Engg.
Sant Longowal Institute of Engg. And Technology,
Sangrur, Punjab.
29
xv
About the Theme of the conference
Automatic Identification refers to the methods of automatically identifying objects, collecting data about
them and entering that data into computer system without human involvement. Technologies considered as a
part of Automatic Identification are Barcodes, Radio Frequency Identification (RFID), Biometrics, Optical
Character Recognition (OCR), Smart cards, Voice Recognition.
Automatic Identification is the process or means of obtaining external data, through analysis of images, sound
and videos. In biometric security systems, capture is the acquisition of or the process of acquiring and
identifying characteristics such as finger image, palm image, facial image, iris print or voice print which
involves audio data and the rest all involves video data. . The technology acts as a base in automated data
collection, identification and analysis systems worldwide.
Biometrics are the oldest form of identification. Dogs have distinctive barks. Humans recognize faces. On
the telephone, your voice identifies you. Your signature identifies you as the person who signed a
contract. Your voiceprint unlocks the door of your house. Your iris scan lets you into the corporate offices.
You are your own key.
In order to be useful, biometrics must be stored in a database. You can verify a signature only if you
recognize it. To solve this problem, banks keep signature cards. The user signs his/her name on a card
when he/she opens the account, and the bank can verify the signature against the stored signature to
ensure that the check was signed by the user.
There is a variety of different biometrics. In addition to the three mentioned above, there are hand
geometry, fingerprints, iris scans, DNA, typing patterns, signature geometry (not just the look of the
signature, but the pen pressure, signature speed, etc.). The technologies are different, some are more
reliable, and they'll all improve with time.
Biometrics are hard to forge: it's hard to put a false fingerprint on your finger. The moral is that biometrics
work well only if the verifier can verify two things: one, that the biometric came from the person at the
time of verification, and two, that the biometric matches the master biometric on file. If the system can't
do that, it can't work. Biometrics are unique identifiers, but they are not secrets. You leave your
fingerprints on everything you touch, and your iris patterns can be observed anywhere you look.
RFID has found its importance in a wide range of markets including livestock identification and Automated
Vehicle Identification (AVI) systems because of its capability to track moving objects. These automated
wireless Automatic Identification Systems are effective in manufacturing environments where barcode labels
could not survive.
RFID is more advanced technology because data can be scanned from a distance, which is not possible in a
barcode. Barcodes are also not different for each item lying in a store but all the RFID stickers will have
different id numbers, which makes it altogether unique. Some airlines are also experimenting with baggage
tracking systems as well with these tags.
xvi
Biometrics : A perspective from the Security Industry
Mr. M. Anand
General Manager
Godrej & Boyce Ltd.
Biometrics: Derived from “bios” meaning “life” and “metron” meaning measure, is literally the field of
identifying and recognizing human beings through their biological attributes.
Biometrics, finds applications in mainly three areas:
u Forensics
u Security
u Human Comfort
As the subject is a wide ranging one, we cannot do much justice to it on a two page writeup. Hence, we will
restrict our discussion to an overview of the various biometric solutions that find applications in the Security
Domain and the challenges posed.
Why Biometric Solution?
A biometric characteristic cannot be forgotten, lost, stolen or copied. Thus it directly resolves the challenges
faced with regards to verifying the person's claim concerning their identity. A password or ID card does not
provide the same level of security as they can be easily compromised.
Pre-requisites for a successful implementation:
For a successful implementation of a biometric solution - in a commercial application - it should meet the
following criteria :
1. Error rates : FAR and FRR: The system should be robust enough to yield very low False Acceptance
Rates and False Rejection Rates.
2. Scalability : The system should be scalable easily for the user to benefit from the implementation.
3. User Privacy : Privacy concerns should be kept in mind while designing the system as biometric
characteristics, by its definition, generally cannot be changed.
4.System Security : The system should be secured so that the confidentiality of biometric data are
maintained with proper authorizations clearly defined.
5.Social acceptance: The system should be accepted by the users. Biometric characteristics like Fingerprint
/ Iris Scanning etc can generate concerns in the minds of the users regarding their effects and hence
confidence should be built up.
6.System Cost: For a successful implementation, the perceived benefits should not only outweigh the costs
incurred, but also the costs should be low enough to encourage trial in the first place
7.Ease of use: And finally, the system should be easy to use.
Biometric Characteristics :
The biometric characteristic that is chosen for deployment can be amongst the following available currently.
Depending on the need, the system can be designed to be a multi modal type requiring authentications by
multiple characteristics or unimodal.
Fingerprint : This is the most common solution under implementation today. The costs are considerably
low due to the high volumes.
Face Recognition: After fingerprint, this is the second most widely deployed solution. Main application
areas are access control, time and attendance applications and law enforcement.
Iris scanning : A solution with very low levels of FAR and FRR, it is gaining increasing levels of acceptance
for indoor applications.
1
Voice Recognition : This helps in identifying persons for remote access, through telephones or internet.
Vein Pattern : It is an emerging technology and we will have to see how it is going to evolve
Hand Geometry: Proven systems for difficult environment, especially in a 1: 1 environment.
Attributes of Biometric Characteristics :
While choosing the suitable characteristic to be considered, the following requirements should be kept in
mind.
Uniqueness : The characteristic should be unique to the individual.
Universality : The characteristic should be available across the entire human population.
Permanence : The characteristic should not change with time.
Measurable : It should be measurable repeatedly with a high level of accuracy.
User friendly : Should be easy to use.
Application areas :
Biometric solutions find application predominantly in the physical access control domain. Solutions have
been tailored to meet the demanding needs of handling thousands of passengers in a busy airport, critical
nuclear installations etc etc.
The benefits that accrue to its users are mainly
Compliance to national and international standards
Risk Mitigation with respect to data thefts
Accurate time data logging and avoidance of buddy punching
Challenges for deployment :
Lack of an uniform industrywide standard : The main roadblock is the lack of uniform international
standards for comparison of the effectiveness of the different biometric technologies / manufacturers. The
claims of FAR / FRR, enrolment time, speed of authentication are critical parameters which a deployer
considers when deciding on the technology and vendor. However, due to lack of uniform evaluation
protocols, one is left to sift and verify the the claims and counterclaims of the various competing technologies.
Costs of deployment : Another significant barrier for widespread acceptance is that the cost of deploying
a biometric solution is still quite high. With technological breakthroughs the costs can come down
significantly in the near future leading to a quantum jump in the usage.
References :
Resources from L1 Identify solutions and IIT Kanpur
Name : M. Anand
Designation : General Manager ( Mktg ) - Electronic Security Solutions
Organisation : Godrej & Boyce Mfg Co Ltd, Mumbai
Education: B Tech in Chemical Engineering and MBA from IIM Bangalore in
Marketing
An industry expert with 14 years of rich experience in designing and deploying
electronic security solutions across a wide gamut of industries, corporates, defence establishments and other
mission critical research centres.
2
RFID (Radio Frequency Identification) : A Tool for Library Security
Prof. N. Barse
Librarian
K. G, Joshi College of Arts, Thane
There have been many technological developments in the last 5o years. The Information Communication
Technology (ICT) revolution in the last decade has had a drastic and far-reaching impact on all aspects of
professional endeavor particularly in the knowledge and information sector. Due to rapid developments in
ICT and its application to the library work libraries have evolve from traditional libraries to automated
libraries, electronic libraries, hybrid libraries, digital libraries and now we are talking about virtual library.
The new technologies have not only transformed the shape of modern libraries but also created many exciting
possibilities and opportunities.
Advances in ICT have increased capabilities such as high resolution capture devices, dramatic increase of
digital storage media, explosive growth of internet and www, fast processing power and reducing cost of
computer, high band with networks and increasing number of security system in digital domain. The digital
developments include infrastructure, acceptability, access, restriction, readability, standardization,
authentication, preservation, copyright, user interface etc. Due to these developments and its impact on
libraries the library collection has undergone many changes. Now days along with books, CDs, VCDs,
DVDs Subscribed online collection (databases) have become important part of the library collection. Most of
the expensive information sources are available in different medias viz. print, electronic and online format.
New technologies have always been of interest for libraries both for increasing the quality of service and for
improving administrative efficiency. Libraries are applying various technological tools for book
identification, self-checkout, automated circulation, theft control and inventory control. These applications
can lead to significant saving in labour; enhance customer services, lower book theft-rate and provide of
updated library statistics for administrative purpose.
Security of library collection has been always a major concern for libraries. Security problems, such as theft
and other kind of misbehavior have become challenges today. Libraries have been using various means for
preventive security measures. Bar code is fundamental technology for library circulation and other
operations.
RFID Technology
Radio Frequency Identification (RFID) is a generic term for technologies that use radio waves. It is an
automatic identification method, relying on storing and retrieving data using electronic tags.
RFID (Radio Frequency IDentification) is the latest technology to be used in library theft detection systems.
Unlike EM (Electro-Mechanical) and RF (Radio Frequency) systems, which have been used in libraries for
decades, RFID-based systems move beyond security to become tracking systems that combines security with
more efficient tracking of materials throughout the library including easier and faster charge and discharge,
inventory (stock checking), and materials handling. RFID is a combination of radio-frequency-based
technology and microchip technology. RFID technology is useful in taking inventory, finding missing items
and identifying misfiled items.
Radio Frequency Identification (RFID) technology collects, uses, stores, and broadcasts data. Components of
RFID systems include tags, tag readers, computer hardware (such as servers and security gates) and RFIDspecific software such as RFID system administration programs, inventory software, etc.
RFID technology can enable efficient and ergonomic inventory, security, and circulation operations in
libraries. In addition like other technologies that enable self-checkout of library materials, RFID can enhance
individual privacy by allowing users to checkout materials without relying on library staff.
3
RFID tag used in a library is a small object, like an adhesive sticker, that can be attaché to or incorporated into
a product. RFID tags contain microchip that is attached to an antenna. The chip and the antenna together are
called an RFID transponder or an RFID tag. The antenna enables the chip to transmit the identification
information to a reader. The reader converts the radio waves reflected back from the RFID tag, into digital
information that can be passed on to a computer that can make use of it.
RFID Use in Libraries
RFID technology can be used in libraries with its greatest advantages for reducing material handling time,
accurate inventory, better collection management, and avoidance routine repetitive tasks of library
professionals. The saved time can be used for quality enhancement in reader oriented and value added
services to its users.
Advantages RFID Technology
® RFID labels or tags can be attached to divergent media in library such as CDs, DVDs, and other printed
materials.
® It can help in speedy circulation because multiple items can be read at a time.
® It combines both material identification and security in one single tag or transponder, thus saving cost and
time.
® Tags are read/write, provides facility for encoding and decoding.
® Tags are durable and can last lifetime of the item they identify.
® RFID provides superior security coverage. It prevents unauthorized removal of library materials from the
premises.
® Stock management such as managing materials on the shelves, finding items that are missing and
identifying miss-shelved items.
® Can be useful in providing improved patron services by spending minimal time on circulation operations
allows library staff to assist users.
® Library items identification and security bit is comibined into a single tag, there by eliminating the need to
attach additional security strip, hence minimizing labeling time and its associated cost.
® Security bit is automatically deactivated and reactivated as materials checked out and in; hence no
separate security procedures are required.
® RFID emit a radio signal that can be picked up from a remote site. Line of sight that involved in barcodes is
not necessary for capturing the item data.
Disadvantages of RFID
High cost : The major disadvantage of RFID technology is its cost.
Chances of removal of tags : RFID tags are typically affixed to the inside back cover and are exposed
for removal.
Depends on other technology like sensor readers for security
While the short-range readers used for circulation charge and discharge and inventorying appear to read the
4
tags 100 percent of the time, the performance of the exit sensors is more problematic. They must read tags at
up to twice the distance of the other readers. The author knows of no library that has done a before and after
inventory to determine the loss rate when RFID is used for security. Lacking data, one can only conjecture that
the performance of exist sensors is better when the antennae on the tags are larger.
Some issues of privacy concerns
Since this technology is capable of tacking the location of an item carrying the tag there is debate on the
privacy related issues
Conclusion :
RFID technology is very useful for all the libraries of the modern world where security of library material is
serious issue. It has the capability of making our professional activities with ease and convenience. Use of
RFID will certainly increase the functional capabilities of the library professionals because they will get more
time for planning and providing better services to the users. It can play significant role in bringing user
satisfaction, and convenience. If we keep aside the cost factor involved in this technology it the best suitable
technology for library management.
Bibliography :
Sahu, Ashok Kumar. 2008. information management in new millennium. New Delhi, Ess.
Sukula, Shiva. 2008. Information technology: bridge to the wired virtuality. New Delhi, Ess.
Carr, Reg. 2007. The academic research library in a decade of change. Oxford, Chandos Publishing.
Webliography :
http://www.ala.org/a/2008la/aboutala/offices/oif/statementspols/otherpolicies/rfidguidelines.cfm
accessed on 14/11/2008
http://www.rfid-library.com/?gclid=CISqoZ3Q55cCFcIupAodxThsDg accessed on 16/11/2008
http://www.rfidgazette.org/libraries/ Accessed on 25/12/2008
5
Multimodel Recognition Implementating Fuzzy Decision Trees
+
M. Gopal and Susmita Deb*
Abstract
It has been found that crisp decision about genuine and imposter match is not efficient enough to implement for
practical purpose. More ever single model is not reliable for genetic factors, aging, environmental or
occupational reasons.
So multi model fuzzy decision based recognition is suggested to take into account the uncertainities. It is
observed that accuracy of recognition is high. Three models are considered: Iris, Fingerprint and Palmprint.
Keywords
Fuzzy Decision Tree, Iris, Fingerprint, Palm Print, FAR, FRR, Matching Score, WEF.
Fuzzy Decision Trees
Fingerprint Recognition
A statistical property known as information gain is
used to measure the worth of attribute and based on
worth of attribute a decision tree is grown.
A fingerprint is the pattern of ridges and valleys on
the surface of a fingertip.
Fingerprint images is converted to gray-scale
image, then normalized, oriented and converted to
binary image. The image is then thinned and
complemented. Only ridge ending and ridge
bifurcation minutiae are considered[2]. Their
position and orientation is found out. Spur, border
minutiae, hole, bridges are eliminated[5].
Fuzzy Id3 is the extension of this classical decision
tree to extract knowledge in uncertain classification
problems which applies fuzzy set theory to
determine the structure of the tree[1].
Fuzzy C-mean clustering algorithm is applied.
Example:
Ridge Oriented Finger Print
0
Scanned Finger Print
Normalised Finger Print
50
100
For 3 attributes A 1, A 2,A3.
150
200
250
300
0
Filtered Image of Finger-Print
50
100
150
200
250
Thinned Image of Finger Print
Post-processed Finger-print
Finger Print with Minutiae (Ridge End & Ridge Bifurcation)
________________________________________________________
+
*
Professor, Indian Institute of Technology, Delhi
PhD fellow, Indian Institute of Technology, Delhi
6
Processed Image of Finger-print
300
350
400
Fingerprintt under test is translated, ro
otated, scaled
and matcheed with a fingeerprint image in
n database.
Histogram of genuine
g
finger-print
Histogram of imp
poster finger-print
80
5000
70
4500
4000
60
3500
50
3000
40
2500
2000
30
1500
20
1000
10
0
500
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0
1
0
0.1
0.2
Matching Score
0.3
0..4
0.5
0.6
0.7
0.8
Matching
g Score
False Rejection Ratio & False Acceptance Ratio plot
Wavelet Energy Function (WEF) is used as feature
vector (4)
100
90
80
FAR& FRR
70
60
The wavelet energy in horizontal, vertical and
diagonal directions at ith level can be, respectively,
defined as :
50
40
30
FR
RR
20
FAR
10
0
0.1
T
0.2
0.3
0.4
0.5
0.6
0.7
0.8
hreshold
Palm Prin
nt Recognition
n
Principal liines, wrinkles and ridges are the features
in a palm pprint. Principall lines, wrinkles and ridges
can be aanalyzed in low, medium and high
resolution respectively as principal lines are
thickest, w
wrinkles are thinner and ridges are the
thinnest. T
Therefore a multi resolution method is
used to anaalyze the palm print.
Pre-processsing is done to
o capture the same
s
area of
palm.
Where M is the total wavelet decomposition level,
can describe the global details feature of a palm
effectively.
Wavelet Energy Feature
The detail images are divided into SS non-overlap
Blocks equally and then the energy of each block
Are computed. The energies of all the blocks are
used to construct a vector. Finally, the vector is
normalized by the total energy. This normalized
vector is named Wavelet Engery Feature (WEF). If
an image is decomposed to J level, the length of its
WEF is 3SSJ.
a b cd
a.
b.
c.
d.
Captured palm
moothen palm
Sm
Paalm with boundary
Seegmented portion is highlighted.
City block distance between two palm print feature
vector is used as matching criterion.
7
Plot of city block distanc
ce of imposter & genuine
e palm
boundaary) the outer bboundary is thee edge of a circcle
with a radius
r
of ri = 34 + rp.
0.9
9
0.8
8
City Block Distance
0.7
7
0.6
6
Imposter Palm
0.5
5
0.4
4
0.3
3
0.2
2
Genuine Palm
0.1
0
0
50
100
150
200
250
300
350
400
45
50
500
P
Palm
No.
Histogram of im
mposter palm
3000
Histogra
am of genuine palm
Norma
alization
45
2500
40
35
2000
30
25
The no
ormalization prrocess projectss iris region in
nto
a consttant dimension
nal ribbon so that
t
two imag
ges
of the same iris under different conditions hav
ve
charactteristic featuress at the same sp
patial location..
Cartesiian to polar ttransformation is done whicch
projectts the iris diskk to a rectangu
ular region wiith
prefixeed size.
1500
20
1000
15
10
500
5
0
0
0
0.1
0.2
0.3
3
0.4
0.5
0.6
0.7
0.8
0
0.2
0.4
0.6
0.8
1
1.2
1.4
City Block Distance
Cityy Block Distance
False Rejection Ratio & False Acceptance Ratio Plot
45
40
35
FRR & FAR
30
25
FRR
20
15
FAR
10
5
0
0.4
0.45
0.5
0.55
0.6
0.65
T
Threshold
Iris Recognition
Iris is high
hly reliable bio
ometric technollogy because
of its stabiility, and the high
h
degree of variations in
irises betw
ween individ
duals. Iris is prefiltered,
localized, normalized,
n
deenoised, enhanced and then
feature is extracted
e
using
g Wavelet.
De-no
oising and en
nhancement
For sy
ymmetry of iriis half of the iris with som
me
upper and lower boundary is considered is
consideered.
High frequency noise is remov
ved, contrast is
improv
ved with the heelp of histogram
m equalization..
Prefilterin
ng
abcd
a.
b.
c.
d.
Original Image
O
Complemented Image
Fiilling holes
Prreprocessed Im
mage
Featurre extraction
Localization
ure vector connsists of 4-leveel decompositio
on
A featu
coefficients of image via wavelet allong vertical an
nd
ntal direction.
horizon
a two nonIris bounddaries can bee supposed as
concentric circles. We must determin
ne the inner
b
witth their relevan
nt radius and
and outer boundaries
centers. Our method is based on iris localization
using imag
ge morphologiical operators and suitable
threshold (50). After pu
upil edge deteection (inner
8
Recognition
Hamming distance between two iris codes is used
as matching criterion.
HD =
1
N
∑
( XOR (CodeA(i ), CodeB (i )) length ( code )
Histogram of imposter iris
Histogram of genuine iris
4000
35
3500
30
3000
25
2500
20
2000
15
1500
10
0
Conclusion
1000
5
500
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
0.6
0
0.4
0.65
0.45
0.5
0.55
0.6
0.65
0.7
Our work involves application of fuzzy decision
trees in biometric. As every biometric recognition
method suffers from some loaf holes, like synthetic
finger print, palm print can replace original one,
dilation of eye can affect iris recognition but in our
case we have put a check for dilation, if it found to
be dilated, image is rejected. Deciding threshold for
declaration of genuine and imposter is quite a
difficult task. For any threshold point FAR (False
Acceptance Rate) & FRR (False Rejection Rate)
are not ignorable. So we are suggesting a
hybridization of the methods, fuzzyfying the
thresholds and taking decision of “
matching”
or
“
not matching”
using fuzzy decision trees.
References
0.75
Hamming Distance
Hamming Distance
False Rejection Ratio & False Acceptance Ratio plot
50
45
40
FAR & FRR
35
30
25
FAR
20
FRR
15
10
5
0
0.38
0.4
0.42
0.44
0.46
0.48
0.5
0.52
0.54
Threshold
Result
Database
5 different images for iris, palm print, finger print
are collected from 50 persons.
1.
Data 62250 x 3, 2 classes.
2.
Hamming Distance between one iris image of a
person and rest four iris images of the same person
(genuine class), 49X5= 245 iris images of different
persons (imposter class) are found out. Similarly
city block distance and matching score for palm
print and fingerprint are found out. Database is
constructed with these data and specified classes.
3.
Randomly chosen 80% data is used for training and
rest 20% data used for testing.
4.
Randomly uncertainity of data is also considered.
The time taken for growing whole tree is
159.386000 seconds
----------------Testing started--------------ans =The time taken for testing is 1.201000
seconds
leaf_no =
5.
3
Accuracy = 98.2892
9
Guoxiu Liang, A comparative study of
three Decision Tree algorithms: ID3,
Fuzzy ID3 and Probabilistic Fuzzy ID3,
Informatics & Economics, Erasmus
University, Rotterdam, Augustus, 2005.
Anil K. Jain, Arun Ross, and Sharath
Pankanti,: A Tool for Information
Security,
IEEE
transactions
on
information forensics and security, vol. 1,
no. 2, June 2006.
Ahmad N. Al-Raisi, Ali M. Al-Khouri,
Iris recognition and the challenge of
homeland and border control security in
UAE, Abu Dhabi Police GHQ, Ministry of
Interior, Abu Dhabi, United Arab
Emirates, Telematics and Informatics,
June 2006.
Xian-Qian WU, Kuan-Quan Wang, David
Zhang, “
Wavelet Based Palmprint
Recognition”
,
First
International
Conference on Machine Learning and
Cybernetics, Beijing, 4-5 November 2002,
0-7803-7508-4/02/ 2002 IEEE.
Raymond Thai, Fingerprint Image
Enhancement and Minutiae Extraction,
The University of Western Australia,2003
Specialized Global Features for Off-line Signature Recognition
H. B. Kekre
Department of Computer Engineering,
V A Bharadi
Department of Electronics and Telecommunication.
Thadomal Shahani Engineering College, Bandra (West), Mumbai-50.
Hbkekre@yahoo.com, vinu_bharadi@rediffmail.com
verification of signature. We divide the features in
Abstract
two types
In this paper we discuss specialised global features
1. Standard Global features
that can be used for off-line signature recognition.
These features are mainly morphological features.
2. Special features
These features can be used to develop an accurate
We discuss four specialized features they are
off-line signature recognition system. These
1. Signature Trisurface area
features are cluster based features they include
2. Grid & Texture Information features
Walsh coefficients of pixel distribution, Vector
3. Walsh Coefficients of Horizontal and vertical
Quantization based codeword histogram, Grid &
pixel distributions.
texture information features and Geometric centers
4. Vector Quantization based-Codeword Histogram
of a signatures. We present the FAR, FRR achieved
These features are used and tested in development of
by the system using these features.
an offline signature system. In this paper we provide
an overview of this feature set.
INTRODUCTION
Signature verification is an important research area
CONVENTIONAL GLOBAL FEATURES
in the field of authentication of a person [3]. We can
Pre-processing is the first step of any signature
generally distinguish between two different
recognition system, where we normalize the
categories of verification systems: online, for which
signature to make it a binary template. This
the signature signal is captured during the writing
normalized template is then used for feature
process, thus making the dynamic information
extraction. The extracted feature vector is used for
available, and offline for which the signature is
comparison and classification of signatures. The
captured once the writing process is over and thus,
performance of signature recognition system is
only a static image is available. In this paper we deal
greatly influenced by the feature set. The prewith Off-line signature verification System. Here
processing is shown in Fig. 1. Signature on the left is
we discuss a set of parameters that can be used in any
original scanned signature; the signature on the right
of the signature verification system for
is the pre-processed signature.
classification of the signatures.
The steps in pre-processing are as follows
Over the years, many features have been proposed
1. Noise Removal
[7-10] to represent signatures in verification tasks.
2. Intensity Normalization
We distinguish between local features, where one
3. Scaling
feature is extracted for each sample point in the input
4. Thinning
domain, global features, where one feature is
This signature template is then used for feature
extracted for a whole signature, based on all sample
extraction.
points in the input domain, and segmental features,
where the signature is subdivided into segments and
one feature is extracted for each segment.
Here we focus on global features. Signature
Figure 1 Pre-processing of a signature
verification can be considered as a two-class pattern
recognition problem, where the authentic user is a
The conventional features that are considered are as
class and all the forgers are the second class. Feature
follows
selection refers to the process by which descriptors
1. Number of pixels - Total Number of black pixels
(features) extracted from the input-domain data are
in a signature template
selected to provide maximal discrimination
2. Picture height - The height of the signature image
capability between classes.
after horizontal blank spaces removed.
Here we consider only global features for
10
3. Picture width - The width of the image with
vertical horizontal blank spaces removed
4. Maximum horizontal projection - The
horizontal projection histogram is calculated and
the highest value of it is considered as the maximum
horizontal projection.
5. Maximum vertical projection - The vertical
projection of the skeletonized signature image is
calculated. The highest value of the projection
histogram is taken as the maximum vertical
projection.
6. Baseline shift - This is the difference between the
y-coordinate of centre of mass of left and right part.
7. Dominant Angle feature - Dominant angle of the
signature.
8. Signature surface area - here we consider the
modified tri-area feature [26]
Features 1 to 6 are self explanatory, we elaborate
feature 7 & 8 in detail.
Figure 3 : Extraction of the 'trisurface' feature signature
(a) original Signature (b) Normalized Signature (c) trisurface
area calculation
The trisurface area calculation requires area
generation for the signature. This gives the three
components of signature area a1, a2, a3. It was
found that the trisurface area feature gives better
classification rate. The features discussed above as
shown in Table I for the signature shown in fig 3.
Table I
Feature Extracted from signature shown in Fig.3
Sr.
Feature
Extracted Value
No.
1
Number of Pixels
547
2
Picture Width (in pixels)
166
3 Picture Height (in pixels)
137
4 Horizontal max Projections
12
5
Vertical max Projections
15
6 Dominant Angle-normalized
0.694
7
Baseline Shift (in pixels)
47
8
Area 1
0.151325
9
Area 2
0.253030
10
Area 3
0.062878
DOMINANT ANGLE FEATURE
First, the signature image was separated vertically
into two equal parts. The position of the centre of
gravity in each part was calculated (A and B in Fig.
2). The angle between the horizontal axis and the
line obtained by linking the two centres of gravity
was the feature added to convert this angle in the
range of 0 to 1. A right triangle was considered using
the centres of gravity as the extreme points of its
hypotenuse, the row of the first centre of gravity and
the column of the other, intersecting to form the right
angle.
GRID & TEXTURE INFORMATION
FEATURE EXTRACTION
Grid and texture feature provide information
about the distribution of pixels and the distribution
density of the pixels [4]. Texture feature provide
information about the occurrence of specific pattern
in the signature template. These features are not
based on single pixel or whole signature but they are
based on group of pixels or signature segments,
hence these are cluster features. Grid feature gives
information about the pixel density in a segment and
texture feature gives information about the
distribution of specific pixel pattern. These features
are discussed in detail here.
Grid information feature
We have the pre-processed signature template. The
resolution is 200*160 pixels. To extract the grid
information feature from the signature we use the
following algorithm.
Algorithm for grid feature extraction
1. Divide the skeletonized image into 10 X 10 Pixels
blocks. We get total 320 blocks.
Figure 2 : Extraction of the 'Dominant Angle' features
SIGNATURE TRISURFACE FEATURE
The surface area of two visually different signatures
could be the same. For the purpose of increasing the
accuracy of a feature describing the surface area of a
signature, the 'trisurface' [4] feature is implemented,
as an extension, in which the signature is separated
into three equal parts, vertically. Fig. 3 illustrates
this concept. The surface area feature is the surface
covered by the signature, including the holes
contained in it. The number of black pixels in the
surface is counted, and the proportion of the
signature's surface over the total surface of the
image is calculated. This process is used for the three
equal parts of the signature, giving three values
between 0 and 1.
11
2. For each block segment, calculate the area (the
sum of foreground pixels). This gives a grid feature
matrix (gf) of size 20 X 16
3. Find minimum and maximum (min, max) values
for pixels block. Ignore block with no pixels.
4. Normalize the grid feature matrix by replacing
each nonzero element 'e i, j' by
of specific pixel pattern.
To extract the texture feature group, the cooccurrence matrices of the signature image are used.
In a grey-level image, the co-occurrence matrix pd
[i, j] is defined by first specifying a displacement
vector d = (dx, dy) and counting all pairs of pixels
separated by d and having grey level values i and j.
In our case, the signature image is binary and
therefore the co-occurrence matrix is a 2 X 2 matrix
describing the transition of black and white pixels.
Therefore, the co-occurrence matrix Pd [i, j] is
defined as
This gives matrix with all elements within the
range of 0 and 1. The results are normalized so that
the lowest value (for the rectangle with the smallest
number of black pixels) would be zero and the
highest value (for the rectangle with the highest
number of black pixels) would be one.
5. The resulting 320 elements of the matrix (gf) form
the grid feature vector.
A representation of a signature image and the
corresponding grid feature vector is shown in Fig.
5.1. A darker rectangle indicates that for the
corresponding area of the skeletonized image we
had the maximum number of black pixels. On the
contrary, a white rectangle indicates that we had the
smallest number of black pixels.
Where p00 is the number of times that two white
pixels occurs, separated by d. p01 is the number of
times that a combination of a white and a black pixel
occurs, separated by d. p10 is the same as p01. The
element p11 is the number of times that two black
pixels occur, separated by d. The image is divided
into eight rectangular segments (4 X 2). For each
region the P (1, 0), P (1, 1), P (0, 1) and P (-1, 1)
matrices are calculated and the p01 and p11
elements of these matrices are used as texture
features of the signature.
Figure 5 shows the grid feature matrix. These values
are scaled by 100 and converted to integer type for
representation purpose only. While processing the
feature vector we use the normalized values only.
Texture feature [4]
Texture feature gives information about occurrence
Figure 6 Pixel positions while scanning for the displacement vector.
We use this procedure to calculate the texture feature
matrix the signature template is divided in eight
segments as follows,
12
similarity between to sequences of the
coefficients and hence the similarity of the two
signatures.
VECTOR QUANTIZATION BASED-CODEWORD
HISTOGRAM
Next we propose a parameter based on Vector
Quantization of a signature template. We segment
the signature into blocks to form the vectors. These
vectors are represented by codewords from the
codebook. Here the codebook used for mapping the
image block plays very important role. Here the
objective of the vector quantization is not to
compress the image but to classify the signature and
verify the authenticity. Hence we extend the
approach to serve our purpose. We use the codeword
distribution pattern as a characteristic of the
signature.
Figure 7 Pixel positions while scanning for the displacement vector
(representation).
Figure 8 Texture feature matrix for signature shown in Fig 7.
WALSH TRANSFORM OF PIXEL
DISTRIBUTION
Here we propose a novel parameter for
signature recognition. This parameter is derived
from the Pixel Distribution of the signature. This is
shown in Fig. 9.
Figure10 Hadamard Coefficients of Horizontal pixel distribution
(Upper Plot) and their sequenced arrangement (Lower Plot) of the
signature mentioned in Fig3
The frequency of the codewords occurring in the
distribution is calculated and the histogram is plotted
for the codeword group against the number of
occurrences and finally these distributions are
compared using a similarity measure used in [5],
which is based on Euclidian distance.
The normalized Signature template is taken, the
image size considered is 200 x 160 Pixels, this image
is divided into 4 x 4 pixel blocks; each block is
treated as a code vector. Overall 2000 blocks will be
there.
The codebook is having all the 216 codewords
initially then the invalid code vectors are removed
by thinning process as shown in Fig 11., and the
remaining code vectors are sorted as there
appearance in gray code table so that the consecutive
code vectors will have minimal change. Such similar
code vectors are grouped to form a vector groups.
The codeword histogram is generated for given
signature template which will be used for
comparison. This operation is illustrated in Fig. 12.
Figure 9 Signature and its horizontal and vertical pixel
distributions
We use Hadamard transform to the horizontal
pixel distribution points (Hi) and vertical pixel
distribution points (Vi); Hadamard transform is fast
to calculate and gives moderate energy compaction.
This operation gives the horizontal Hadamard
coefficients (HH i ) and vertical Hadamard
coefficients (VHi). We use Kekre's [10] algorithm to
get the Walsh coefficients from the Hadamard
coefficients. This operation yields Walsh
coefficients of the histograms (SHHi, SVHi). The
coefficients are plotted and shown in figure 4.
These coefficients are calculated for the test
signature as well as standard signature and
Euclidian Distance is evaluated to measure the
13
The thinning operation gives 11756 valid
codewords. This is the set of codewords used for
group formation. Next step is to form codeword
groups. To form codeword groups we start with
initial codeword chosen from the set and compare
with the other codewords to find group of
codewords with minimum hamming distance. For
current scenario we have constructed 240 groups of
50 codewords each. We form one extra group for the
codewords which do not participate in any group.
The observed intra group hamming distance is in the
range 2-6 and in extreme case distance such as 8-9
are observed.
The codeword histograms can be compared using
the similarity measure shown in Equation 5.3. For
all these features we have used user specific
thresholds which give better results as compared to
the common thresholds. We have used set of training
signatures containing 8 signatures. The thresholds
are calculated by using training procedure discussed
in [1].
RESULTS
For testing we have used total 800 signatures from
75 users and 150 forgery signatures including
skilled as well as simple forgeries. We have tested
individual modules for each cluster based feature
discussed above and the final system implementing
the feature set together. The final system uses user
specific thresholds and training mechanism
discussed in [1].We have implemented a weighted
comparator based classifier in the final system. The
metrics FAR & FRR [4] are evaluated for each
feature. They are as follows,
Table I
False Acceptance and False Rejection Ratios reported by the
Figure 12 Vector Quantization applied for signature template
Similarity Measure [5]
Given an encoded image having similar
representation as a text document, image features
can be extracted based on codevectors frequency.
The feature vector for signature template I1 and the
feature vector for test signature I2 are given below,
For I1, It is given by I1= {W11, W21, … Wn1} (5.1)
For I2, It is given by I2= {W12 , W22 , … WN2 }
(5.2)
In the histogram model, Wij = Fij , where Fij is the
frequency of group Ci appearing in Ij .Thus, the
feature vectors I1 and I2 are the codeword
histograms. The similarity measure is defined as
Sr.
No.
1
2
3
4
5
6
Feature
Geometric Centres
Walsh Coefficients
Vector Histogram
Grid Feature
Texture Feature
Final System
FAR
FRR
16%
40%
12%
8%
14%
5.70%
14%
42%
22%
12%
20%
8.80%
CONCLUSION
In this paper we have discussed specialized
global features. These feature set can be used to
develop an Off-Line signature recognition system.
We have developed an Off-line signature
recognition system based on these features. The
system uses all these features and user specific
thresholds. The system is having accuracy of 94.3%.
These features are easy to implement and can be
explored in a deeper approach to achieve higher
recognition rates.
(5.3)
Where the distance function is
(5.4)
We get the codeword histogram as shown in Fig. 13
REFERENCE
[1] B. Majhi, Y S Reddy, D Prasanna Babu,
“NovelFeatures for Off-line Signature verification”,
International Journal of Computers,
Communications & Control Vol. I (2006), No. 1, pp.
Figure 13 Codeword Histogram for Signature shown in Fig. 9
14
17-24.
[2] H. Lei, S. Palla, V. Govindrajalu, “ER2- an
Intuitive similarity measure for online signature
verification”, State University of New York at
Buffalo, Amherst NY 14260, USA, Research paper
by CEDAR
[3] A. Kholmatov , “Biometric Identity Verification
Using On-Line & Off-Line Signature
Verification”,Master of Science thesis, Sabanci
University,S 2003
[4] H. Baltzakis, N. Papamarkos, A new signature
verification technique based on a two-stage neural
network classifier, Engineering Applications of
Artificial Intelligence 14 (2001), 0952- 1976/01/$ PII: S0952
[5] L. Zhu, A. Rao & A. Zhang, “Theory of
Keyblock-based Image retrieval”, ACM Journal,
Volume V, No. N, March 2002, PP 1-32
[6] H B Kekre, T. Sarode, “Clustering Algorithm for
codebook generation using Vector Quantization”,
Proceedings of National Conference on Image
Processing 2005, TSEC, Mumbai.
[7] A. K. Jain, A. Ross, S. Prabhakar, “An
Introduction to Biometric Recognition”, IEEE
Transactions On Circuits And Systems For Video
Technology, Vol. 14, No. 1, January 2004
[8] J. Soldek, Shmerko, et.al, “Image Analysis and
Pattern Recognition in Biometric Technologies”,
Proceedings International Conference on the
Biometrics: Fraud Prevention, Enhanced Service,
Las Vegas, Nevada, USA, 1997, pp. 270-286
[9] H. Dullink, B. van Daalen, J. Nijhuis, L.
Spaanenburg, H. Zuidhof, “Implementing a DSP
Kernel for Online Dynamic Handwritten Signature
Verification using the TMS320 DSP Family”,
EFRIE, France December 1995 SPRA304.
[10] J. K. Solanki, “Image Processing using fast
orthogonal transform”, PhD Thesis Submitted to
IIT Mumbai, 1978, pp. 30, 31
15
Fingerprint Recognition in DCT Domain
1
M. P. Dale, 2S. N. Dharwadkar, 3M. A. Joshi
1,2 MES’s College of Engineering, Pune
3 College of Engineering, Pune
used to compare the corresponding pixels of the
fingerprint images in many of rotating and shifting.
In ridge feature based matching, the feature of ridge
is extracted from the gray scale image and
represented by other parameters. Those parameters,
depending on tools used in feature extraction
process are then used for matching fingerprints. The
GABOR filter-based method is an example of ridge
feature based matching [4]. Using the similar
approach transforms can be employed for feature
extraction. Ridge structure of fingerprint can be
considered as oriented texture pattern having
dominant spatial frequency. Transform coefficient
of an image tend themselves as new feature, which
have the ability to represent the regularity,
complexity and some texture feature of an image.
This paper present the fingerprint recognition
method based on Discrete Cosine Transform (DCT)
approach. This approach is applied first on cropped
images in experiment I while the same approach is
applied on resized images in experiment II. With
proper selection of DCT based features, a superior
performance is achieved compared to the Discrete
Wavelet Transform (DWT) based method proposed
in [3]. To demonstrate the efficiency of our
approach, the performance in terms of % Genuine
Acceptance Rate (%GAR) andprocessing time
required for training and testing between DCT and
DWT are compared. Section 2 describes the
experimental settings and database creation. The
DCT and DWT basics are discussed section 3.
Feature description using DCT is briefed in section
4. Results and discussions are presented in detail in
section 5. Conclusions are finally drawn in section 6.
Abstract
Fingerprint verification is one of the most reliable
personal identification methods and it plays a very
important role in forensic and civilian applications.
This paper proposes DCT features for fingerprint
matching. In the feature extraction process,
fingerprint image is cropped and quartered around
core point. Each quartered sub-image is DCT
transformed and standard deviation of properly
grouped transformed coefficients is calculated as
feature. Such standard deviations from four subimages gives feature final vector Euclidean and
Canberra distance are used for final matching in
identification mode. The results are represented in
terms of % Genuine Acceptance Rate (GAR). The
training and testing time is also compared.
Results are further improved using resizing of
fingerprint images.
1. Introduction
Fingerprint recognition is considered one of the
most reliable and mature technologies and has been
extensively used in personal identification.
Fingerprints are graphical patterns of ridges and
valleys on the surface of fingertips, the formation of
which is determined during the first seven months of
fetal development. One kind of widely used feature
is minutiae, which is usually defined as the ridge
ending and the ridge bifurcation. Most of the studies
of fingerprint recognition are based on minutiae
features [1]. Despite its simplicity and efficiency in
storage, minutiae-based representation has its
drawbacks in practical usage. As a kind of local
feature, the minutiae is difficult to be extracted
robustly due to various factors such as large
displacement, different pressure, noise etc. Further
the spurious minutiae will degrade the performance
seriously [2]. Also in the matching step of minutiae
based algorithm, minutiae need to be revised
according to the template to conquer the
translational and the rotational distortion of the
fingerprint images which is time consuming and
make the fingerprint identification system
unpractical on large fingerprint database.
Other approaches used for fingerprint matching are
correlation based and ridge feature based matching.
The correlation based matching is a straight method
2. Experimental settings and database creation.
To extract transform (DCT) features from
fingerprint, gray- scale fingerprint images from 72
individuals were collected using Biodesk
fingerprint optical scanner from NITGEN company
having the resolution of 500dpi. The subjects mainly
consisted of student and staff volunteers from our
Institute in the age group of 20 to 50 and 40% are
male. The size of the original image grabbed is
248x292 as shown in Figure 1. Eight images per
person are collected at different timings. Out of
those 8 images k (2, 4 or 6) images are used for
16
training and all the images in database (72 X 8 =
576) are used for testing.
like Discrete Fourier Transform (DFT), DCT can be
factored as Kronecker products of several smaller
sized matrices, which leads to fast algorithms for
their implementation.
The DCT is similar to DFT, since it decomposes a
signal into a series of harmonic cosine functions
given by
where 0 ≤ u ≤ M-1, 0 ≤ v ≤ N-1. The DCT is real,
orthogonal, fast and separable transform. It has
excellent energy compaction for highly correlated
data. DCT convert images from time (spatial)
domain to frequency domain to decorrelate pixels
while DWT provides good localization both in time
and spatial frequency domain. The DWT refers to a
class of transformations that differ not only in
transformation kernel employed but also in the
fundamental nature of those functions and the way
in which they are applied. The various transforms
are related by the fact that their expansion functions
are “small waves” kernels (hence the name
wavelets). The DWT is identical to hierarchical sub
band system where the sub bands are
logarithmically spaced in frequency and represent
octave band decomposition.
(A)
(B)
4. Feature description using DCT
In experiment I, the algorithm first find out the core
point of the image using slope technique described
in paper [5]. Image is then cropped around the core
point into size of 64x64. This gives translation
invariance and also considers that area of fingerprint
around the core point where most of the information
is concentrated. Figure 2(a) shows original image
and Figure 2(b) shows the cropped image around
core point which is further divided into four sub non
overlappingsub-images as shown.
(c)
Figure 1 (a). Original Fingerprint Image of size 248x292 1
(b) 8 images cropped around core point. 1 (c) 8 resized
images.
Considering most of the information of fingerprint
is near the core point, in experiment I we first
detected the core point with slope technique [5] and
then fingerprint image is cropped around it of size
64X64. Figure 1(b) shows 8 image samples of one
person cropped around core point. For experiment II
instead of using the core point and cropping, we
used resizing of whole fingerprint image to a size of
64X64. Figure 1(c) shows an example of resized
images of one person. This not only reduced large
time of core point detection but also improved
%GAR.
3. DCT and DWT basics
The term image transform refers to a class of unitary
matrices used for representing images. Images are
expanded in terms of a discrete set of basis arrays
called basis images[8]. Energy conservation, energy
compaction, decorrelation are the important
properties of these transforms. Image transforms
(a)
(b)
Figure 2. (a) Fingerprint image (b) Four 32x32 pixel
sub−images, cropped and quartered at its centre.
The 2-D transform (DCT) is applied on each
subimage separately [7]. The DCT transformed
17
coefficients are now grouped into different
frequency bands (blocks) as shown in Figure 3. For
eachnumbered block the standard deviation is
calculated using formulas given in equation (2) and
stored as feature vector.
Table 1 (a) and (b) shows the % GAR based on
above mentioned two distances respectively. The
DWT results are obtained according to paper [3]
with 3 level decomposition of each sub-image and
taking standard deviation as feature of each
decomposed detail band. Table 2 shows the time
required for training with 2 images per person and
testing time in identification mode with all (576)
images without considering the core point detection
time. Timings are calculated. In our test machine
used is Pentium R D CPU 2.8 GHz.
where p(i.j) is the transformed value in (i,j) for any
block or sub-band of size N x N. Such features are
calculated from four images and hence form a total
of 36 features will form feature vector which is used
in enrollment as well as in matching phase. These
are stored as float values so take 144 bytes to store
the feature vector for one person.
Table 1(a). %GAR for cropped images with
Euclidean distance (TS − Training set)
Method
DCT
Wavelet-db9
Figure 3. Grouping of bands in DCT
%GAR
TS=2
TS=4
72.39 78.99
66.14 75.52
TS=6
82.81
79.16
Table 1(b). %GAR for cropped images with
Canberra distance (TS− Training set)
It is found from fingerprint image that, as image is
traversed towards core point from outer periphery,
the distance between ridges becomes less and area
surrounding the core is denser. From recognition
point of view, it is better to extract more information
of ridges and core area. In the DCT domain this area
corresponds to top right, bottom left and bottom
right corner, that is nothing but mid and high
frequency region. In this area, by considering only 9
blocks per sub-image and their corresponding
standard deviations as feature vector the recognition
rate is calculated.
Method
DCT
Wavelet-db9
%GAR
TS=2
TS=4
78.12 82.29
75
79.16
TS=6
85.93
84.89
Table 2. Training and testing time.
Method
DCT
Wavelet-db9
5. Results and discussion
During the training phase 72 person images are
given as input. The training set used per person may
be 2, 4 or 6 fingerprint images per person. Because
more than one image is used in training phase the
final feature vector is taken as average standard
deviation. Euclidean distance and Canberra
distance given by (3) and (4) is used to find out the
minimum distance match between two dimensional
feature vectors of database image (x) and query
image (y).
Time in seconds
Training
Testing
5.76
6.6
18
53
As seen from the tables, the DCT results are better in
terms of recognition rate (%GAR) as well as training
and testing time as compared to wavelet. Further
Canberra distance gives better results as compared
to Euclidean results [9].
The major drawback of above method is time
consuming task of core point detection and still the
%GAR is required to be improved. In experiment II,
to try out the algorithm for whole image at the same
time to restrict the size of image to 64X64, resizing
of images is done as shown in Figure 1(c). After
resizing of images the above three algorithms are
applied as it is. Results in terms of %GAR are shown
in Table 3 (a) and (b).
18
Table 3(a). %GAR for resized images with
Euclidean distance (TS− Training set)
Method
DCT
Wavelet-db9
%GAR
TS=2
TS=4
71.35 86.11
70.83 82.98
presented, it is concluded that DCT coefficients and
their proper combination (bands) results in very
informative description for fingerprint pattern
recognition. The experimental results also shows
that, with method discussed applied on resized
images, a higher recognition rate together with a
lower complexity could be achieved in fingerprint
matching system as compared to the wavelet
features. The results are obtained without any
preprocessing (enhancement) of images. In future,
the algorithms will be implemented on standard
database and investigation will be carried out in
terms of the possibilities of feature fusion to
improve %GAR.
TS=6
90.75
89.93
Table 3(a). %GAR for resized images with
Canberra distance (TS− Training set)
Method
DCT
Wavelet-db9
%GAR
TS=2
TS=4
73.78 86.28
74.62 85.76
TS=6
93.75
91.66
7. References
The obtained results after resizing clearly shows
improvements in %GAR with DCT again giving the
better results. The second major advantage here is in
terms of time. The core point detection
approximately takes 10 sec for one fingerprint
image while resizing takes only 16 msec. Further
DCT implemented with FFT approach takes 4 times
less time in training and testing phase. From the
results it is obvious that the DCT coefficients have
more variations than the DWT, which provides a
higher resolution for fingerprint features matching.
Hence the informative features derived from DCT
coefficients are then more distinguishable than that
from the DWT coefficients. According to study , the
impact of image quality degradation on
performance of on fingerprint matching systems
shows that this type of ridge based methods perform
better than minutiae based method on lesser quality
images.
[1] D. Maltoni, D. Maio, A.K. Jain and S. Prabhakar,
Handbook of Fingerprint Recognition, Springer, New
York, 2003.
[2] Dingrui Wan and Jie Zhou, “Fingerprint Recognition
Using Model-Based density Map,” IEEE Transaction on
Image Processing, Vol. 15, No. 6, 2006
[3] M. Tico, E. Immonen, P. Ramo, P. Kuosmanen and J.
Saarinen, “Fingerprint Recognition Using Wavelet
Features”, The 2001 IEEE International Symposium on
Circuits and Systems, Vol.2, pp. 21-24, 2001.
[4] Muhammad Umer Munir and Muhammad Younas
Javed, “Fingerprint Matching using Gabor Filters”,
National Conference on Emerging Technologies 2004.
[5] Anil K. Jain, Salil Prabhakar, Lin Hong and Sharad
Pankanti, “Filterbank based Fingerprint Matching”, IEEE
Transaction on Image Processing, Vol. 9, No. 5, May
2000.
[6] S. Tachphetpiboon and T. Amornraksa, “Applying FFT
Features for Fingerprint Matching,” IEEE Trans., 2006.
[7] T. Amornraksa and S. Tachaphetpiboon, “Fingeprint
Recognition using DCT Feature,” Electronic Letters, Vol.
42, No. 9, 2006.
[8] A. Jain, Fundamentals of Digital Image Processing,
Englewood Cliffs, NJ: Prentice-Hall, 1989.
[9] Mahesh Kokare, B. N. Chatterji and P. K. Biswas,
6. Conclusions and Future work
From the exhaustive experiments conducted with
actually captured fingerprint images and the results
19
AUTOMATIC IDENTIFICATION TECHNIQUES
Frequency Ranges and Radio Licensing Regulations
Mr. Rahul Kulkarni
Senior Manager
Siemens
Mrs. Kirti S. Agashe
HOD, Industrial Electronics Dept.
V.P.M.'s Polytechnic,
Chendani Bunder Road, Thane 400601, Email: kirtiagashe@yahoo.com
Abstract
Automatic Identification and Data Capture (AIDC) is the method of automatically identifying objects, collecting
data and storing that data directly into computer systems for further processing and analysis. These
technologies typically include bar code, OCR, Biometrics, Smart card, Voice Recognition and Radio
Frequency Identification (RFID). These all techniques are commonly referred to as AIDC “Automatic
Identification,” “Auto-ID,” or "Automatic Data Capture."
Radio frequency identification (RFID) is comparatively a new AIDC technology, first developed in 1980's.
RFID has found its importance in a wide range of applications because of its capability to track moving objects.
These automated wireless AIDC systems are effective in environments where barcode labels could not
survive.
This paper highlights on the various techniques of Auto Identification in brief and about the Frequency ranges
and radio licensing regulations for the widely accepted Auto Identification methods and other communication
applications implemented in the same range.
Automatic identification procedures (Auto-ID) have become very popular in many service industries,
purchasing and distribution, industry, manufacturing companies and material flow systems. Automatic
identification provides the information about people, animals, goods and products in transit.
The most important automatic identification techniques include
 Barcode Systems
 Optical Character Recognition (OCR)
 Biometric Procedures: Voice Identification & Fingerprinting Procedures (Dactyloscopy)
 Smart Cards
 Radio Frequency Identification (RFID)
Identification with extremely cheap barcode labels was a revolution in identification system and has proved to
be successful over last 20 years. The barcode is a binary code comprising a field of bars and gaps arranged in a
parallel configuration. They are arranged in a predetermined pattern and represent data elements related to an
associated symbol. The sequence, comprising of wide and narrow bars and gaps, can be interpreted
numerically and alphanumerically. The optical laser scanner read it. Approximately ten different barcode
types are in use at present. The most popular barcode is the EAN code (European Article Number), which
was designed to fulfill the requirements of the grocery industry in 1976.
The constraints of this technique are their low storage capacity and the inability in reprogramming.
Optical character recognition (OCR) technique was first used in the 1960s. Special fonts were
developed for this application so that they could be read by human being as well as by machines,
automatically. The most important advantage of OCR systems is the high density of information and the
possibility of reading data visually in an emergency.
Presently OCR is found to be useful in production, service, and administrative fields, banks for the
registration of cheques.
However, OCR technique has failed to become universally applicable due to the high price and the
complicated reading procedures.
20
Biometric procedures
Biometrics is defined as the science of counting and measurement procedures involving living beings.
Biometry is the general term for all procedures that identify people by comparing unmistakable and
physical characteristics of an individual. Such as, fingerprints and handprints, voice identification and, less
commonly used, retina (or iris) identification.
Voice identification
Recently, specialized systems have become available to identify individuals using voice verification (voice
recognition). In such systems, the user voice is stored in a digital form in computer and the identity of that
speaker is preserved. Next time the user is identified with the existing reference file.
Fingerprinting procedures (dactyloscopy)
For the identification of criminals, fingerprinting procedures are used since the early twentieth century. This
process applies the comparison of papillae and dermal ridges of the fingertips. This can be obtained from the
finger itself, as well as, from objects that the individual has touched.
When fingerprinting technique is used for personal identification, for example entrance procedures, the
fingertip is placed upon a special reader. The system forms a data record from the pattern it has read and
compares it with an existing stored reference pattern. Latest fingerprint ID systems require less than half a
second to identify a fingerprint.
Smart cards
A smart card is an electronic data storage system with additional computing facilities. It is incorporated into
a plastic card, which resembles a credit card. In 1984,the first smart card in the form of prepaid telephone
smart card was launched.
Smart cards are placed in a reader .The contact springs in the reader make a galvanic connection to the contact
surfaces of the smart card. The smart card is supplied with power and a clock pulse from the reader via the
contact surfaces. Data transfer between the reader and the card is via a bi-directional serial interface (I/O port).
The primary advantage of the smart card is the fact that the data stored on it can be protected against
unauthorized (read) access and manipulation. Smart cards provide all required information especially for
financial transactions and make it simpler, safer and Cheaper. But disadvantage of contact-based smart cards
is its vulnerability of the contacts to wear, corrosion and dirt. Its maintenance is expensive and shows
malfunctioning sometimes. In addition, readers that are accessible to the public cannot be protected against
vandalism.
Based upon the internal functionality two types of cards are available which are the memory card and the
microprocessor card.
Memory cards
In memory cards usually an EEPROM is used. Sometimes simple security algorithms are also incorporated.
These memory cards are very cost effective.
For this reason, memory cards are predominantly used in price sensitive, large-scale applications.
Microprocessor cards
Microprocessor cards contain a microprocessor along with a segmented memory (ROM, RAM , EEPROM
segments).
The mask programmed ROM stores an operating system for the microprocessor and is provided by the
manufacturer and can not be altered. The EEPROM usually contains application data and application
oriented programs.The RAM is the volatile temporary working memory.
21
Microprocessor cards are very flexible and are primarily used in security sensitive applications, for GSM
mobile phones and the new electronic cash cards.
RFID systems
The mechanical contact used in the smart card is not very practical. It is desired to have a contact less transfer
of data between the data-carrying device and its reader . In the ideal case, the power required to operate the
electronic data-carrying device should also be transferred from the reader using contact less technology. Such
procedures for the transfer of power and data,are offered in the contact less ID systems which are called RFID
systems (Radio Frequency Identification).
An RFID system comprises of two components
The transponder, which is located on the object tobe identified. Data is stored on this electronic datacarrying device and it normally consists of a coupling element and an electronic microchip
The interrogator or reader, which is the data capture device and is also called as the reader.
A reader typically contains a radio frequency module comprising of a transmitter and a receiver, a control unit
and a coupling element to the transponder. Many times readers are provided with an additional interface (RS
232, RS 485, etc.) to enable their communication with PCs and related systems.
Data
Contactless
RFID
Reader
Clock
Energy
Data carrier =
Transponder
Coupling Element
Application
(Coil, microwave
Antenna)
Frequency Ranges and Radio Licensing Regulations
5.1 Frequency Ranges Used
RFID systems generate and radiate electromagnetic waves hence they are legally classified as radio systems.
It is thus essential to ensure that under no circumstances, the function of other radio services such as radio,
television, mobile radio services (including police, security services, industry) marine radio services,
aeronautical radio services and mobile telephones should be disturbed by the operation of RFID systems.
Therefore the radio services are specifically reserved for industrial, scientific or medical applications. These
frequencies are classified worldwide as ISM frequency ranges (Industrial Scientific Medical), and they can
also be used for RFID applications. In addition to ISM frequencies, the entire frequency range below 135 kHz
(in North and South America) and <400 kHz (in Japan) is also suggested. This is because it is possible to
work with high magnetic field strengths in this range and particularly with inductively coupled RFID
systems. Thus the most important frequency ranges for RFID systems are 0135 kHz, and the ISM frequencies
around 6.78, 13.56 Hz, 27.125 MHz, 40.68 MHz, 433.92 MHz, 869.0MHz, 915.0MHz (not in Europe), 2.45
GHz, 5.8 GHz and 24.125 GHz.
Frequency range 9135kHz
The range below 135 kHz is mainly used by other radio services, as it has not been reserved as an ISM
frequency range. These frequencies can propagate within 1000 km continuously, at a low technical cost.
Aeronautical and marine navigational radio services operate in this range.
22
Frequency range 6.78MHz
The range 6.7656.795MHz is used as the short wave frequencies. The daytime propagation of these
frequencies range is up to a few 100 km and during nighttime transcontinental propagation is possible. This
frequency range is used by a wide range of radio services, like broadcasting, weather reports, aeronautical
radio services and press agencies.
Frequency range 13.56MHz
The range 13.55313.567MHz is situated in the middle of the short wavelength range. This range allows the
transcontinental connections during day time and is used by a wide variety of radio services like press
agencies and telecommunications (PTP) and remote control systems, radio equipment and pagers.
Frequency range 27.125MHz
The frequency range 26.56527.405 is allocated to CB radio across the entire European continent, in the USA
and in Canada.
The ISM range between 26.957 and 27.283MHz is situated approximately in the middle of the CB radio
range. Applications like inductive radio system, diathermic apparatus medical
application, high frequency welding equipment, remote controlled models and pagers are possible in this
range.
Frequency range 40.680MHz
The range 40.66040.700 MHz is situated at the lower end of the VHF range. In this range the propagation of
waves is limited to the ground wave. This frequency range is used by mobile commercial radio systems for
forestry, motorway management and by television broadcasting .The major ISM applications that are
operated in this range are telemetry and remote control applications.
Frequency range 433.920MHz
The frequency range 430.000440.000MHz is allotted to amateur radio services. Radio amateurs use this
range for voice as well as data transmission and for communication via relay radio stations and home-built
space satellites.
The ISM range 433.050434.790MHz is located approximately in the middle of the amateur radio band and is
used by a wide range of ISM applications.
Applications like backscatter (RFID) systems, baby intercoms, telemetry transmitters, cordless headphones,
unregistered LPD walkie-talkies for short range radio, vehicle central locking and many other applications
are possible in this frequency range.
Frequency range 869.0MHz
In 1997 in Europe, the frequency range of 868870 MHz was passed for Short Range Devices (SRDs) in and is
thus available for RFID applications in the 43 member states of CEPT.
Frequency range 869.0MHz
In 1997 in Europe, the frequency range of 868870 MHz was passed for Short Range Devices (SRDs) in and is
thus available for RFID applications in the 43 member states of CEPT.
Frequency range 915.0MHz
This frequency range is not available for ISMapplications in Europe but in USA and Australia the frequency
ranges 888889MHz and 902928 MHz are available and are used by backscatter (RFID) systems.
Frequency range 2.45GHz
The ISM range 2.4002.4835 GHz partially overlaps with the frequency ranges used by amateur radio and
radiolocation services. For this UHF frequency range, buildings and other obstacles behave as good LAN
23
reflectors and damp electromagnetic waves at transmission. Applications like backscatter (RFID) systems,
telemetry transmitters and PC LAN systems for the wireless networking of PCs are carried out in this
frequency range.
Frequency range 5.8GHz
The ISM range 5.7255.875 GHz partially overlaps with the frequency ranges used by amateur radio and
radiolocation services.
Typical ISM applications for this frequency range include movement sensors, used as door openers or contact
less toilet flushing and also for backscatter (RFID) systems.
Frequency range 24.125GHz
The ISM range 24.0024.25 GHz overlaps partially with the frequency ranges used by amateur radio and radio
location services and also by earth resources services via satellite.
This frequency range is used by movement sensors, directional radio systems for data transmission.
References:
 En.wikipedia.org/wiki/automatic_identification and_data_capture-29k
 en.wikipedia.org/wiki/Automatic_Identification _System
 en.wikipedia.org/wiki/RFID -*
24
RFID, Its Applications and Integrating RFID in WSN
Anuradha Kamble (Lecturer)
SIES Institute of Technology, Nerul, Navi Mumbai
Anita Diwakar (Lecturer)
V. P. M’s Polytechnic, Thane
ABSTRACT :
Radio Frequency Identification (RFID) and Wireless Sensor Network (WSN) are two important components
of pervasive computing, since both technologies can be used for coupling the physical and the virtual world.
Radio-frequency identification (RFID) is an automatic identification method, relying on storing and remotely
retrieving data using devices called RFID tags or transponders. The technology requires some extent of
cooperation of an RFID reader and an RFID tag. Here in this paper, section 1 deal with the basics of RFID
systems and its applications. Section 2 discusses architecture of wireless senor network and section 3 defines
the way to integrate RFID in wireless sensor network.
Keywords : RFID, Wireless sensor network, wireless network, smart nodes etc.
INTRODUCTION :
It is widely believed that the next revolution in computing technology will be that the widespread small
wireless computing and communication devices will integrate seamlessly into daily life. We can therefore
expect in the near future lots of devices to grow by multiple orders of magnitude such as tags, sensors etc.
They gather information about the current environment, which means sensing and processing information.
In recent years, automatic identification procedures (Auto-ID) have become very popular in many service
industries, purchasing and distribution logistics, manufacturing systems etc. Automatic identification
procedures exist to provide information about people, animals, goods and products in transit.
Currently available barcode labels that triggered a revolution in identification systems, some considerable
time ago, are being found to be inadequate in an increasing number of cases. Barcodes are extremely cheap
but problem is with their low storage capacity and the fact that they can not be reprogrammed.
The technically optimal solution would be storage of data in a silicon chip. The most common form of
electronic data carrying device in use in everyday life is the smart card based on contact field. However
mechanical contact used in smart card is often impractical. A contact less transfer of data is more flexible,
which suggests use of RFID (Radio frequency identification.) [2]
1. RFID Overview :
An RFID tag is an object that can be applied to or incorporated into a product, animal or person for purpose of
identification and tracking using radio waves.
RFID devices have three primary elements: a chip, an antenna, and a reader. A fourth important part is a
database where information about tagged objects is stored.
The chip, usually made of silicon, contains information about the item to which it is attached. Chips used by
retailers and manufactures to identify consumer goods may contain 'Electronic Product code'. The EPC is the
RFID equivalent of the familiar Universal Product Code or the barcode. Barcodes must be optically scanned
and contain only generic product information .By contrast , EPC chips are encrypted with a unique product
code that identifies the individual product to which the tag is attached and can be read using radio frequency.
The antenna, attached to the chip is responsible for transmitting information from the chip to the reader,
using radio waves. The chip and the antenna combination is refereed to as a transponder or, more commonly
as a tag.
25
1
The reader, or scanning device, also has its own antenna, which it uses to communicate with the tag.
Reading tags refers to the communication between the tag and the reader via radio waves operating at a certain
frequency. In contrast to bar codes, one of the RFID's principal distinctions is tags and readers can
communicate with each other without being in each other's line of sight.
The database, or other back end logistics system, stores information about RFID tagged objects.
Operating Procedures :
RFID systems operate according to one of two basic procedures :
In half duplex procedure (HDX), data transfer from transponder to the reader alternates with data transfer
from reader to transponder. At frequency below 30MHz this is most often used with load modulation
procedure.
In full duplex procedure (FDX), data transfer from transponder to reader takes place at same time as data
transfer from reader to transponder. However both procedures have in common the fact that transfer of energy
from reader to transponder is continuous i.e. it is independent of direction of data flow. In sequential systems,
transfer of energy from transponder to reader takes place for limited period of time only. Data transfer from
transponder to reader occurs in pauses between power supply to transponder.[2]
RFID Applications :
 Inventory Control : Currently, most material tracking systems employ two-dimensional barcodes that
must be close to and within the "line of sight" of the barcode reader. This requires manual scanning or a
conveyor-like process to position the barcode and scanner. Barcodes can run the risk of getting wet or
scratched due to mishandling or a harsh environment, which often prevents accurate reading by the scanner.
Manual intervention is labor intensive, costly, and error-prone. In addition, scheduled scanning or manual
methods cannot ensure the inventory remains up-to-date, due to oversights, errors, and internal shrinkage.
RFID solution, inventory can be updated in real time without product movement, scanning or human
involvement. Our fully automated system allows inventory status to be determined and shipping & receiving
documents to be generated automatically. The system could also trigger automatic orders for products that are
low in inventory.
 Container and pallet tracking : RFID active tags can be programmed with contents and assigned
locations and then placed on containers and pallets that are stored in a warehouse. Additional information can
be collected and added to the RFID tags as the pallets move through the warehouse. The tracking system can
identify unscheduled movement, so managers and security can be alerted to possible theft.This system can
also reduce theft and other forms of inventory shrinkage by immediately alerting the operations manager
when a product is moved from its assigned area.
 Airports and high Security : RFID access control and personnel tracking and location systems can
help to assure the security of restricted areas in airports, such as flight lines, baggage handling areas, customs,
employee lounges, and other sensitive areas. These techniques could also be applied at maritime ports,
railway stations and passenger bus terminals. RFID systems can be used to track employee and passengers in
real time. Unlike other access control systems, RFID solutions are completely hands free so they have
minimal interference with the busy work schedules and flow of employees and passengers.
 Equipment tracking in hospitals : RFID system can be used to track patients, doctors and
expensive equipments in hospitals in real time. RFID tags can be attached to the ID bracelets of all patients, or
just patients requiring special attention, so their location can be tracked continuously. RFID technology can
also provide an electronic link for wirelessly communicating patient data. An instant assessment of critical
equipment and personnel locations is also possible through RFID technology.
 Parking lot control : RFID technology can provide independent, non-stop systems for security,
parking, and access control. RFID technology provides businesses and communities with hands-free control
to ensure only authorized vehicles have entry. The system can also provide access data for administering
periodic access charges or parking fees. [1][2]
 Manufacturing Lines : Manufacturers can track and record in-process assembly information into the
RFID tag as an item progresses along the line. For example, as features are added to a personal computer
26
assembly, they could be recorded on the tag. In this case, the tag would keep a current "inventory" of the PC's
contents. The tag information could later be read to produce a shipping list and invoice. The tag could also
remain with the item for later use by field personnel during installation and maintenance.
2. Wireless Sensor Network:
Advances in hardware and wireless network technologies have created low-cost, low-power, multifunctional
miniature sensor devices. These devices make up hundreds or thousands of ad hoc tiny sensor nodes spread
across a geographical area. These sensor nodes collaborate among themselves to establish a sensing network.
A sensor network that can provide access to information anytime, anywhere by collecting, processing,
analyzing and disseminating data. Thus, the network actively participates in creating a smart environment.
The figure shows the complexity of wireless sensor networks, which generally consist of a data acquisition
network and a data distribution network, monitored and controlled by a management center. [3]
Figure 3: Wireless Sensor Network [2]
Data is sensed by the sensors placed at various applications. They transmit time series of the sensed
phenomenon to the central nodes where computations are performed and data are fused. A sensor network is
composed of a large number of sensor nodes, which are densely deployed either inside the phenomenon or
very close to it. The position of sensor nodes need not be engineered or pre-determined. This allows random
deployment in inaccessible terrains or disaster relief operations. On the other hand, this also means that sensor
network protocols and algorithms must possess self-organizing capabilities. Another unique feature of sensor
networks is the cooperative effort of sensor nodes. Sensor nodes are fitted with an on-board processor. Instead
of sending the raw data to the nodes responsible for the fusion, sensor nodes use their processing abilities to
locally carry out simple computations and transmit only the required and partially processed data.[4]
3. Integration of RFID Reader and WSN Base Station :
One trend of the development of RFID is integrating
it into network. RFID network is very mature now
such as Real Time Locating System (RTLS), which
implies us to integrating sensor nodes into RFID to
get more environment information we need. A mix of
tags and sensor nodes are deployed in detected area.
Smart stations gather information from tags and
sensor nodes then transmit it to local host PC or
remote LAN. Here RFID and WSN information can
be integrated in the base station, which will be more
intelligent. For example, WSN data triggers RFID
reader for certain unusual event. The new system will
be composed of three classes of devices. The first
Figure 4:
class is that of wireless devices with no serious power
27
Heterogeneous network architecture [1]
class is that of wireless devices with no serious power constraints named as smart stations. The device will
contain an RFID reader, a 32- bit microprocessor for local data processing and a network connection. They
are nearly identified with the wired devices but use wireless connections to the backbone network for more
convenient deployment.
 Distributed Smart Node :
As no network stack is embedded into the reader in a RFID system at present, the reader can only be operated
passively and all of its behaviors are controlled by local control system. Its very big volume also makes it
difficult to move around. Moreover, the position of antennas of an RFID reader must be computed carefully to
cover all the tags in range and not to conflict with other antennas or readers. All of these disadvantages limit
the applications of RFID. If functions of a reader are cut short, an RFID reader might get much smaller, less
expensive and easy to deploy. We propose a new smart node containing less functional reader. For now there
seems to be no counterpart in RFID systems for this device.
 Smart node :
The smart node contains three parts: sensing part which makes use of kinds of sensors to detect interested
physical scenario, reading part which reads fewer tags comparing with a normal RFID reader, and radio
transceiver which transporting sensed data. Smart nodes read fewer tags and can be deployed densely as selforganizing WSN. Smart nodes run autonomously and translate data information to the sink node. The
gathered information is transmitted through multi-hops.
Figure 5: Smart Node Platform [4]
As information of tags in the same area is similar, it can be compressed with simple and high effective data
compressing methods in each smart node. Consequently, flexible communication protocol is necessary.
Presently, ZigBee protocol is the best candidate for the proposed architecture.
Conclusion :
RFID is an efficient technique for numerous applications, and combining RFID with WSN gives better
results. Here in this paper we propose the network architecture, which has consequent features as different
functional nodes integrated RFID tags or readers with sensor nodes. It consists smart stations mixed RFID
readers and WSN base stations. Base stations are complicated and costly and its very big volume also makes it
difficult to move around. If there is something wrong with a smart station the whole system will break up and
therefore reliability of the system is decreased.
References :
[1] 'Integrating RFID in WSN' by lei zhang and lei wang
[2] RFID handbook by kluas Finkinzeller
[3] 'Survey on wireless sensor network devices' by Marco Augusto M.Vieria and Diogenes 0-7803-79373/03-2003 IEEE
[4] 'Wireless integrated senor network' by G.J. Pottie and W.J. Kaiser Communication of the ACM May 2000
/ Vol. 43, No. 5 51
[5] www.rfidjournal.com
28
A Skin Color Segmentation Approach towards Face Detection
Ashwani Kumar
Department of Electrical and Instrumentation Engineering,
Sant Longowal Institute of Engineering and Technology,
SLIET, Longowal, Sangrur, Punjab, India
Email: ashwani_75@yahoo.com
In contrast to template matching, the models (or
templates) are learned from a set of training images
which should capture the representative variability
of facial appearance [4]. These learned models are
then used for detection. These methods are designed
mainly for face detection.
Abstract :
The field of computer vision involves developing
techniques to detect objects in a digital image. One
of the main applications of the object detection is
human face detection. In this paper, skin color
segmentation approach has been discussed in face
detection system. It is mostly used as a first
approximate localization and segmentation of faces
in the camera image, in order to reduce the search
area for other more precise and computationally
expensive facial feature detection methods. The
advantage of using skin color for segmentation is
that it is invariant of orientation of the human face.
This method can be used for online face detection.
Keywords : face detection; segmentation; edge
detection.
2. Color Segmentation
Detection of skin color in color images is a very
popular and useful technique for face detection [5].
Many techniques have reported for locating skin
color regions in the input image. While the input
color image is typically in the RGB format, these
techniques usually use color components in the
color space, such as the HSV or YIQ formats [6].
That is because RGB components are subject to the
lighting conditions thus the face detection may fail if
the lighting condition changes. Among many color
spaces, YCbCr components have been used. In the
YCbCr color space, the luminance information is
contained in Y component and the chrominance
information is in Cb and Cr. Therefore, the luminance
information can be easily de-embedded [7]. The
RGB components were converted to the YCbCr
components using the following formula. In the skin
color detection process, each pixel was classified as
skin or non-skin based on its color components. The
detection window for skin color was determined
based on the mean and standard deviation of Cb and
Cr component, obtained using 32 training faces in 3
input images. The color segmentation has been
applied to a training image [8]. Some non-skin
objects are inevitably observed in the result as their
colors fall into the skin color space.
1. Introduction
To detect faces from a single intensity or color
image, single image detection methods can be
classified into four categories.
1.1. Knowledge-based methods.
These rule-based methods encode human
knowledge of what constitutes a typical face[1].
Usually, the rules capture the relationships between
facial features. These methods are designed mainly
for face localization.
1.2. Feature invariant approaches.
These algorithms aim to find structural features that
exist even when the pose, viewpoint, or lighting
conditions vary, and then use these to locate faces.
These methods are designed mainly for face
localization [2].
Y = 0.299R + 0.587G + 0.114B
Cb = - 0.169R - 0.332G + 0.500B
Cr = 0.500R - 0.419G - 0.081B
1.3. Template matching methods.
Several standard patterns of a face are stored to
describe the face as a whole or the facial features
separately. The correlations between an input image
and the stored patterns are computed for detection
[3]. These methods have been used for both face
localization and detection.
3. Image Segmentation
The next step is to separate the image blobs in the
color filtered binary image into individual regions.
The process consists of three steps. The first step is
to fill up black isolated holes and to remove white
1.4. Appearance-based methods.
29
isolated regions which are smaller than the
minimum face area in training images [9]. The
threshold, 140 pixels is set conservatively. The
filtered image followed by initial erosion only
leaves the white regions with reasonable areas.
Secondly, to separate some integrated regions into
individual faces, the Roberts Cross Edge detection
algorithm is used. The Roberts Cross Operator
performs a simple, quick to compute, 2-D spatial
gradient measurement on an image. It thus
highlights regions of high spatial gradients that
often correspond to edges [10]. The highlighted
region is converted into black lines and eroded to
connect crossly separated pixels. Finally, the
previous images are integrated into one binary
image and relatively small black and white areas are
removed. The difference between this process and
the initial small area elimination is that the edges
connected to black areas remain even after filtering.
And those edges play important roles as boundaries
between face areas after erosion. The final binary
images and some candidate spots that are compared
with the representative face templates.
in the image and solid background. It is having the
problem when the color of the background is similar
to the human skin color neighboring skin like can
also combined with other skin regions which can
change the shape if the area not to depict the actual
object. Labeling of regions in the image helps in
eliminating the area which is smaller than the
required value. Here the face region of the image is
detected first and then eye and mouth are located for
those regions to confirm that the region is a face
region.
5. Conclusion and Future Scope of work
This work could be expanded and can be used for the
face recognition in real time application. Additional
work would be to find the distances between the
eyes and to calculate the area of the triangle made by
joining the centers of two eyes and mouth. Then
matching the distance between eyes with the
database images. This can filtered out images whose
distance matches with the distance of input image.
Then, the area of the triangle can be matched with
the filtered out images. And these steps can be
further implemented from the database with the
input image.
4. Results and Discussion
The Image with large areas excluding small areas is
shown in Figure 1 and Image after Median Filtering
is shown in Figure 2.
References
1. Ming Husan Yang Narendra Ahuja "Face Detection
Using a Mixing factor Analyzers" University of Illinoisat
Urbana Campaign, Urbana IL 61801
2. Chiunhsiun Lin, Kuo-Chin Fan "Human Face Detection
using Geometric Triangle Relationship" National center
University, Chung-Li,Taiwan. 32054 R.O.C
3. Son Lam Phung, Douglas Chai And Abdesselam
Bouzerdoum "Skin Color Based Face Detection" 100
JoondalUp Drive, Joondalu.Western Australia 6027
4. Rein-Lein Hsu "Face Detection in Color Images"
Mohamed Abdel-Mottaleb, IEEE and Anil K.Jain IEEE
5. Zhongfei Zhang Rohini K. Srihari Aibing Rao "Face
Detection and its Application in intelligent and Focused
Image Retrieval" Binghamton N.Y. 13902, Buffalo
NewYork
6. Rong Xiao, Ming-Jing Li, and Hong-Jiang Zhang
"Robust Multiose Face Detection in Images" Vol 14, No 1, January 2004 Retrieval
7. G. Yang and T. S. Huang, “Human Face Detection in
Complex Background,” Pattern Recognition, vol. 27, no. 1,
pp. 53-63, 1994.
8. C. Kotropoulos and I. Pitas, “Rule-Based Face Detection
in Frontal Views,” Proc. Int'l Conf. Acoustics, Speech and
Signal Processing, vol. 4, pp. 2537-2540, 1997.
9. D. Chetverikov and A. Lerch, “Multiresolution Face
Detection,” Theoretical Foundations of Computer Vision,
vol. 69, pp. 131-140, 1993.
10. Face detection by Inseong Kim,Joon Hyung Shim, and
Jinkyu Yang.
Figure 1. Image with large areas excluding small areas
Figure 2. Image after Median filtering
Human face detection is a first step in the field of
face recognition. Human face detection has many
applications in routine security checks passport for
driving license validation etc. Additional techniques
to remove or smooth out noise in the image were
required, median filter was used for that purpose.
This algorithm finds the instances of front faces in
color images. This works very well with one person
30
VPM CalUniversity Alliance Agreement
Dr. VN BRIMS has entered into educational collaboration with California University of
Technology, USA, (CalUniversity) for conducting value added, professional courses namely, Master of
Business Administration (MBA) and Doctor of Business Administration (DBA).
The CalUniversity, incorporated in USA, is a distance education Institution of higher learning
dedicated to the study of Business Administration and Management promoting quality learning, critical
thinking, and the discovery of new knowledge for the benefit of diverse business communities. There is a
formal procedure for screening the aspirants to be selected for admission for CalUniversity courses. The
institute will provide lecturers, facilitators and state-of-the-art facilities. CalUniversity will also send their
lecturers to VPM as guest lecturers. The MBA (CalUniversity) instruction includes amongst other things,
class contact hours, self-study and tutorials/facilitators sessions. Candidates will get the benefit of high
quality lectures and tutorials directly from the University using teleconferencing & videoconferencing
systems.
The various MBA programs covered under this agreement are as follows : 1.
Master of Business Administration Program with the following emphasis options
a.
International Management and Marketing (IMM)
b.
Information Systems and Knowledge Management (ISKM)
c.
Healthcare Management and Leadership (HCML)
d.
Banking and Finance (BF)
e.
Organizational Development and Human Resource Management (OHRM)
f.
Project and Quality Management (PQM)
2.
Doctor of Business Administration Program with the following emphasis options
a.
Global Business and Leadership (GBL)
b.
Entrepreneurship and Business Management (EBM)
c.
Healthcare Management and Leadership (HCML)
d.
Information System and Enterprise Resource Management (ISERM)
Centre For Foreign Languages
Understanding the need and importance of students to learn foreign languages, Dr. VN BRIMS has started a
Centre for Foreign Languages. The languages offered are Japanese, French, German. There are regular
batches and weekend batches and as of now there are 40 students being trained in various languages. There is
also a separate library in the institute premises for foreign languages.
Seminar on Consolidation-The New Business Mantra
Dr. VN BRIMS is organizing a seminar on 14th February 2009 on the topic “Consolidation-The New Business
Mantra”. The seminar seeks to address the multi-farious issues arising from the new business mantras viz.
Consolidation. The coveted list of speakers for the workshop includes stalwarts like Dr. Vishnu Kanhere, Dr.
Keshab Nandy, Dr. Guruprasad Murthy etc.
31
EIGHT ANNUAL NATIONAL CONFERENCE
'Industrial Safety Practices for
Peace, Productivity and Profits
10th October 2009”
The main objective of VPM is to disseminate knowledge to the next generation. Keeping this in mind, we are
associating this activity with other institutes in and around Thane for maximum participation. Experts from
the following industries would share the practices followed keeping the above theme in mind.

Agriculture

Pharmaceutical/Food

Power

Infrastructure

Chemicals

Engineering

RefineriesAll the above Industries are covered under MSBTE syllabus. At least 3-4 papers will be from
our Alumni. We also look forward for your active interest and make the conference a grand success.
V.P.M.'s Polytechnic has set a trend in conducting annual conferences on subjects of tropical interest. “An
Advanced Diploma in Industrial Safety” curriculum is conducted at our Institute for the last 6 years. This
course is for industrial working personnel who are experienced minimum of 2 years after their Degree or
Diploma in Engineering and Technology or Science graduate. Even some of the students have reached senior
management cadre. Thus this conference is exclusively dedicated to the students, by the students and for the
students of safety.
The current industrial slow down demands, increased productivity, higher efficiency at low cost. In addition
there is a paradigm shift from “Economics” to “Environment”. Today if you can achieve all the above but
environment is polluted, it is not acceptable. Hence we have selected the topic.
Prof. D.K. Nayak
Convener/Principal
V.P.M.'s Polytechnic, Thane
Prof. V. S. Bhakre
Co-convener
Prof. - Industry Institute Interaction
32