A systematic approach for feature extraction in fingerprint images

Fingerprint image enhancement and feature extraction are the most important stages in fingerprint verification. The robustness of the system depends entirely upon its ability to enhance bad quality images and reliably extract minutiae from it. In this paper we propose two original algorithms for the extraction of global and local features in fingerprint images. A global feature extraction and enhancement algorithm based on Fourier analysis that is simultaneously able to extract local ridge orientation, ridge frequency and ridge quality measure is proposed. A novel algorithm for minutiae feature extraction and post processing based on chain coded contour following is also presented.