Protected Biometric Identification with Multiple Finger Vein
Swati K. Choudhary, Ameya K. Naik · 2022 2nd Asian Conference on Innovation in Technology (ASIANCON) · 2022
The access control or gaining permission to do a particular activity without claiming your identity is an identification process which involves one is to all matchings. The proposed multi-instance biometric identification utilizes multiple instances of finger vein to identify a person with secured templates. Common pre-processing and texture-based feature extraction is applied to two finger vein instances followed by feature fusion with dimensional reduction keeping the discriminative features to further simplify the system design making it a real time application. Secured templates are created out of fused finger vein features subjected to Gaussian random projection-based Index-of-Max hashing. The hashed templates are not only secured but also forms compact integer vector requiring low storage space and matching time. Use of multiple instances increases the universality of identification and finger vein makes the system contactless. This contactless property of an authentication system highly recommends its use for any infectious situations like covid-19.