Implementation of biometrie smart card using multibiometrics

Abstract

Authentication of user is important for security system. Presently, Single biometrie trait along with knowledge-based methods and token based methods are some of the most popular approaches. Even though these methods have less security approaches for users this project proposes a multimodal biometric approach through which the users can be authenticated using sclera vein and Ear recognition. Smart cards are turning out to be a fundamental piece of the transformation of retailing into electronic commerce. These are very flexible in providing authentication at different level of the bearer and the counterpart. The Blood vein pattern in the sclera, is stable over time and unique for each person. Hence it is well suited for use as a biometric for human identification and authentication. The sclera recognition rate obtained from this project is 90%. Recognizing people by their Ear is relatively efficient form of biometrics. The shape and features of ear is unique for each person and invariant with age. The ear recognition rate obtained is 90% for our own database. The features of sclera and ear is combined using Feature level fusion and the recognition rate obtained is 93.33% In this paper the results obtained from the feature level fusion of sclera and ear from each person is used as passwords for smart card. by this way using of other traits like finger print, Iris, Finger vein whose disadvantages like usage of high cost Iris scanner, Finger print scanner capturing dirt, greases and contamination found in finger which creates hygienic issues can be overcome. Since single scanner is required to capture both sclera and ear makes this paper highly advantage and cost effective. The features obtained from sclera and ear are stored as templates and a unique id is assigned to each template. These information's are stored in RFID Database totally a group of 6 cards are taken in which 4 provides authentication and 2 provides no authentication. With the biometric template stored on the smart card, comparison can be made locally, without the need for connection to a database of biometric identifiers. Since all biometric matching takes place using templates, it is unnecessary to store complete biometric image data on the smart card.