| Fingerprint represents the most secure way to identify individuals because it provides a novel approach to recognize the identity, or verify the claimed identity through a unique, highly reliable and robust physical characteristic or personal trait.Nowadays, many products of fingerprint have been offered, however, because of the disturbance of image noise and non-linear of the skin elasticity, to achieve an ideal AFIS (Automatic Fingerprint Identification System) is still a difficult research problem. So it is an efficient solution to the problem that to find an efficient and fast algorithm processing fingerprints images, and to study the technology of fingerprint characteristic on computer and network.Parts algorithms of the preprocessing, post-processing, minutiae extraction and matching of AFIS based on minutiae is discussed in this paper.In the aspect of preprocessing, an easy segmentation method is implemented through the analysis of fingerprint images from U.are.U Firefly and the histogram of the images. We also improve on the binarization method of fingerprint images based on the 2nd derivative by the orientation and ridge frequency of the images. And then a new thinning algorithm is implemented on 16-pixels-neighbors zone. Which compose the whole preprocessing algorithm.The false minutiae are eliminated for the thinned image in post-processing,which reduce the response-time requirement in minutiae extraction and matching.In the algorithm of minutiae extraction, a vector is for the information of a minutia, and the minutiae of a fingerprint image are composed of vectors. To improve the matching rate of the rotate images, we improve on the algorithm of minutiae extraction and fingerprint matching.C++ Builder implements all algorithms of AFIS on PC. Experiments results show that all algorithms perform well. The whole system is tested on the hardware of DSP, too, which demonstrates that our system meets the given requirement and works well. |
PDF Full Text Request