Relevant Papers:
- Mansukhani, P., S. Tulyakov and V. Govindaraju(2007). Using Support Vector Machines to eliminate False Minutiae Matches during Fingerprint Verification. SPIE Defense and Security Symposium (DSS), Orlando, FL., USA.
Verification of Fingerprint Matching
In the project of fingerprint match verification we assume that some preliminary match was performed and we have a hypothesized matching between minutias of two fingerprints. The goal of the project is to derive a reliable confidence measure of the match. As we notice, the preliminary matching is almost always able to give a correct hypothesis of few minutia correspondences in the genuine match, but the derived confidence score might be lower than that of the impostor match. The idea of the project is that trainable pattern classification algorithms will be able to better separate minutia correspondences in genuine and impostor matches.
Two directions of research are been investigated: minutia based verification and image correlation based verification.