Smart card based face recognition biometric system

Abstract

Currently with improving of the technology of smart card especially in CPU and capacity of memory and also with developing many new algorithms for detecting Biometric parameters and making high-tech new appearing devices, many intensive research work is in progress. This thesis explains the smart-card-based face verification and recognition system. The project starts with the selection of the client specific linear discriminant analysis (CS-LDA) algorithm. The CSLDA algorithm is developed in CAIRO in Universiti Technologi Malaysia. A smart card based face recognition and verification system is proposed in which the feature extraction and decision making is performed on the computer because of some limitation on the smart card. The verification and recognition will perform image processing technique which is captured from user will send to smart card and also save in the database. With using of the webcam one image captured from card holder and with using CSLDA algorithm, some different files come out from this algorithm. Two files will be send and store on the EEPROM of smart card and some other files with this two files save to the computer as well. The smart card which is selected for this project is one of the microprocessor smart card with 256 kb EEPROM memory and the card reader which is used for this work is standard card reader compatible with ISO smart card. With implementing this project and test the results, verification between the data which is written to smart card and the data which is read from smart card is ok, and also for recognition the user from the database with comparison between the ID of user which is given to him/her during the enrollment stage and all the users which is enrolled in the data base, the software can be detect that this user was enrolled or Not and after testing many times the result is ok. The accuracy of the CSLDA algorithm is not so high and for this reason the online verification is not so good. This implementation project has been performed on basis of Microsoft c# language and Microsoft SQL server.