Trimodal biometrics personal authentication systems using improved score level fusion techniques

Abstract

In the recent years there is an increased emphasis on the privacy and security of information stored in various storage media such as personal devices databases corporate servers especially in fields like military security agencies corporate firms and automatic personal identification has become a very important topic Though there exist several high level security mechanisms using biometric verification most of these uses only a single trait of biometric The problem with a unimodal biometric verification system is that since it uses only a single biometric trait it suffers from the disadvantages such as lack of universality interclass variation and sensitivity to attacks which lead to spoofing of the authentication system In order to overcome these shortcomings ultimodal biometric systems are introduced In this thesis various combinations of multimodal authentication systems are implemented and analysed with several matching level fusion techniques A multimodal system using dynamic fingerprint verification system and improved iris segmentation is fused using adaptive rank level fusion echnique The results proved that the highest rank adaptive rank level fusion technique requires least authentication time of 0 54 minutes when compared with borda count and logistic regression fusion schemes newline