Feature extraction is not only a crucial preprocessing step for fingerprint matching, but is also essentially involved as a core component of automated fingerprint identification (AFI) systems, which mainly aims at detecting singular as well as all other fingerprint minutiae points that impart individuality to each fingerprint and differentiate one fingerprint from another. The key question that constantly arises in fingerprint feature extraction is how to achieve high-accuracy minutiae with inherently varied-quality scanned fingerprints. In this paper, an effective method for minutiae-based fingerprint feature extraction is proposed, where high accuracy minutiae data are first obtained via preprocessing mechanisms, such as denoising, binarization, thinning, bifurcation, termination and ridge to valley area. Then, feature extraction is performed by employing the Harris corner detector that looks for large changes in local
contrast to extract reliable minutiae points from a given fingerprint automatically and quickly. Experimental results indicate that the proposed method exhibit good performance in extracting valid minutiae points, even with low-quality fingerprint images.