Multiple-Biometric Authentication Based On Parallel Computing

This paper provides initial results obtained on a multimodal biometric system that uses face,fingerprint and hand geometry features for biometric verification purposes. Our experimentsindicate that the sum rule performs better than the decision tree and linear discriminantclassifiers. The benefits of multibiometrics may become even more evident in the case of alarger database of users. We are, therefore, in the process of collecting data corresponding tothree biometric indicators – iris, face and signature – from a larger user set.Biometric system is defined as "automated methods of verifying or recognizing the identityof a living person on the basis of some physiological characteristic", "or some aspect ofbehavior"Ross and Jain's experiments in 2003 indicate that the sum rule performs better than thedecision tree and linear discriminant classifiers. User verification systems that use a singlebiometric indicator often have to contend with noisy sensor data, restricted degrees of freedom,non-universality of the biometric trait and unacceptable error rates.Attempting to improve the performance of individual matchers in such situations may notprove to be effective because of these inherent problems.Multibiometric systems seek to alleviate some of these drawbacks by providing multipleevidences of the same identity. These systems help achieve an increase in performance that maynot be possible using a single biometric indicator. Further, multibiometric systems provideanti-spoofing measures by making it difficult for an intruder to spoof multiple biometric traitssimultaneously. However, an effective fusion scheme is necessary to combine the informationpresented by multiple domain experts.This dissertation addresses the problem of information fusion in biometric verificationsystems by combining information at the matching score level. Experimental results oncombining three biometric modalities (face, iris, and signature) are presented.