In this paper, a full algorithm is proposed for fingerprint identification： First, original fingerprint images is normalized. Second, segmentation is carried out in order to differentiate between valid area and invalid area. Orientation image is estimated, the fin- gerprint images is filtered using orientation filtering. Next the enhanced image is binaried with dynamic threshold, the fingerprint image is thinned by improved OPTA thinning algorithm. Then the minutiae data is extracted from the thinned image. At last, point matching is used for ...
