Strategy to extract reliable minutia points for fingerprint recognition

Fingerprint recognitions is one of the most popular biometric identification methods. However, most of fingerprint recognition algorithms seen in the literature have the problem of creating false minutia points or not reliably identifying the valid minutia points because of low image quality. This paper proposes a robust minutiae based approach for fingerprint recognition. The image is first enhanced using multiple thresholding method. Then after binarization, minutia points are extracted using a 3×3 window as described in the paper. We propose a simple strategy to obtain valid minutiae points. For each point under test, the proposed strategy examines the four neighboring points to determine if the current point is a valid bifurcation point. The proposed strategy is tested on a number of images of publicly available FVC2006 database. Results indicate that the proposed strategy filters the false minutiae points reliably.