The detection of singular points(core and delta) is very important for fingerprint classification and some fingerprint matching based on reference points. This paper propose a new algorithm for extracting the singular points using the combination of curvature and orientation field of fingerprint images, which provide a more accurate description of characteristic of singular points. The effectiveness of our algorithm has been shown by the experiments implemented in the public databases, DB2 and DB3, FVC2002.
展开▼