Highly reliable biometric methods for Person authentication in access Control applications

Abstract

Biometric system is essentially a pattern recognition system that verifies newlinea person based on the features extracted from his her physiological or behavioral newlinecharacteristics Generally a biometric system performs four distinct tasks namely newlinepreprocessing of sensor images feature extraction feature matching and decision newlinemaking to provide person authentication Any biometric should possess newlinecharacteristics such as universality uniqueness permanence collectability and newlineacceptability to persons involved However all these characteristics put together newlinemay not be present in every biometric This fact has led to significant research in newlinemulti biometric systems which rely on fusing information from multiple biometric newlinescores similar or diverse biometrics resulting in higher reliability Nowadays newlinethese biometric systems play a significant role in a lot of security related newlineapplications Combined with technological advances the increased concern towards newlinesecurity has created a boost to intelligent person authentication systems newlineIn this regard this research work first proposes procedural improvements newlinefor feature extraction in unimodal biometric authentication systems based on iris newlineretina and ear Secondly different score normalization and fusion schemes by newlinemaking use of multiple snapshots multiple units multiple biometrics are newlineimplemented which are expected to contribute further towards improved intelligent newlineauthentication systems newline newline