An Algorithm for Fingerprint Identification Based on Wavelet Transform and Gabor Feature

This paper presents an algorithm for fingerprint identification based on wavelet transform and Gabor features. Firstly, a center point area of the fingerprint is detected, then the image in this area is decomposed into different sub-images using wavelet transform, finally we extract Gabor features from these sub-images to generate feature vectors for matching. The experiment conducted over the four FVC2004 databases shows that the proposed approach can capture much texture information at different scales and orientations, achieve high recognition rates.