As one of most reliable biometric technology, Automatic fingerprint verification has been apply to many field. We develop some key algorithms of automatic fingerprint identification system (AFIS), main contexts are as following:(1) The direction map of fingerprint image is computed accurately by the change of gray on the tangent direction and the dircection of nomal. It brings outs a new method on smooth of the direction map. Experimental results show tht it is feasible for fingerprint processing.(2) We discuss the dynamic thresholding binarization method in fingerprint image preprocessing and present a new algorithm about binarization enhancement.it can fill the little hole on the ridge effectively.(3) The set of original minutiae is directly extracted from thinned fingerprint images. Then combining with the information of local ridge direction, special algorithms are designed with respect to various noises to delete pseudo minutiae from original minutiae set.(4) An algorithm based on ridge fitting for seeking pairs of reference minutiae is presented .It use the algorithm of error diffused by level in the stage of matching minutiae .(5) A system combined with password and fingerprint is proposed.Experimental results show it is more secure and mord accurate than customary system.All the algorithm in this thesis have been carry out and test in the PC. The result show that the function of the system has been improved greatly . |