Enhanced Fuzzy Feature Match Algorithm for Mehndi Fingerprints

Abstract

The performance of biometric system is degraded by the distortions occurred in finger print image acquisition. This paper focuses on nonlinear distortions occurred due to �Mehndi / Heena drawn on the palm/fingers. The present invention is to detect and rectify such distortions using feedback paradigm. If image is of good quality, there is no need to renovate features. So, quality of whole image is checked by generating exponential similarity distribution. Quality of local region is checked by the ridge continuity map and ridge clarity map. Then, we check whether feedback is needed or not. The desired features such as ridge structure, minutiae point, orientation, etc. are renovated using feedback paradigm. Feedback is taken from top K matched template fingerprints registered in the database. Fuzzy logic handles uncertainties and imperfections in images. For matching, we have proposed the Enhanced Fuzzy Feature Match (EFFM) for estimating triangular feature set of distance between minutiae, orientation angle of minutiae, angle between the direction of minutiae points, angle between the interior bisector of triangle and the direction of minutiae, and a minutiae type. The proposed algorithm incorporates an additional parameter minutiae type that assists to improve accuracy of matching algorithm. The experimentation on 300 Mehndi fingerprints acquired using Secugen fingerprint scanner is conducted. The results positively support EEFM for its efficiency and reliability to handle distorted fingerprints matching.