Minimize Search Time through Gender Classification from Multimodal Biometrics

Abstract

Multimodal Biometrics is an emerging domain that integrates two or more biometric features to overcome certain limitations of using them individually. In this paper, a Multimodal Biometric System integrating Fingerprint, Palmprint and Hand geometry is proposed for identification / verification process, which is essential for a security access of information or region sectors. Though multimodal system results in good match rate, searching time for a match is long and incremental depending on comparison numbers. There is more scope for refinement in the search time. With this motive, this paper discusses a robust suggestion of extracting gender information from the multiple source samples in order to minimize the search time during identification and verification process. Classifying the samples either as male or female in an accurate manner, will help in dividing the entire Biometric database into approximately two halves, one corresponding to male and other pertaining to female, which will be very helpful for real time applications utilized for both identification and verification process of a known and unknown individual. Since various researches have been carried out on determination of gender from fingerprints and palm prints, we suggest the most accurate methodology amongst them for an enhanced identity verification system. Additionally we have also discussed employing Hand Geometry for obtaining gender information to enhance the accuracy further, taking the advantage that these features can be acquired from the same image used for extracting Fingerprint and Palm print features.