Singular Points Of The Fingerprint Image Thinning Algorithm

In this paper, several key problems in the fingerprint identification were discussed. By using the theory and technology of image processing, pattern recognition and computer application, we studied the problems of detecting singular points of fingerprints, classification of fingerprint and fingerprint thinning etc. There are four chapters in this paper. The first chapter is introduction, mainly discussed the fingerprint identification technology and summarized the problem of the fingerprint classification and the fingerprint thinning. The second chapter studied fingerprints singular points detecting and the fingerprint classification. The third chapter is the research in the fingerprint thinning. The fourth chapter is the conclusion. The main contents are as follows:(1) Summarize the fingerprint classification problems.Automatic fingerprint identification is a hotspot in recent years, and fingerprint classification is an important part of the AFIS, it has a significant meaning in application value. In this paper, several fingerprint classification methods were summarized and analyzed. Though there is many works were done in the fingerprint classification in the world, the existing fingerprint classification methods still could not meet the application need in practice because the fingerprint image gather technology's limitation and the complexity of the fingerprint classification. In the fingerprint classification problem, detecting singular points exactly is the kernel. From this meaning, we should go deep into the singular points detecting and the fingerprint classification problem.(2) The research for the fingerprint singular points detecting and the fingerprint classification.There exist two singular points detecting methods—the method based the block image level and the method based the pixel level. But almost all the existing methods have problems of losing true singular points and including fake singular points. In this paper, the two singular points detecting methods were implemented and the performance of the two methods were analyzed. It was found that the detecting method based block image often losing true singular points but almost not detecting fake singular points, and the detecting method based pixel often detecting many fake singular points but almost not losing true singular point. From the comparison, we cansee the two methods have their self-character. Based the character of the two methods, a new method detecting singular points was studied. The method based block image level and the method based pixel level were combined and the new method was proved having a better veracity and reliability than the two original methods from the experiment. The new singular detecting method was used to classify the fingerprints and the classification result in the experiment also proved the validity of the combined singular point detecting method.(3) The research in the fingerprint thinning.Fingerprint thinning is a key part in the automatic fingerprint identification technology. The fingerprint thinning result has an important influence of the AFIS. Fingerprint thinning is making the binary fingerprint image to the one pixel width image. The purpose of the thinning is convenient for detecting minutiae in the following steps. There are several thinning methods used in the fingerprint thinning, but from the thinning result, there are many problems in these thinning methods. In other words, these thinning methods could not meet the need of fingerprint thinning. In this paper, four thinning methods often used were implemented. And the result of these thinning methods were analyzed and evaluated systemic. Then, the method—serial and parallel thinning method was improved on. From the result of the experiment, we can see the thinning result has a good representation. The thinning efficiency is high, and there is not ridges erosion, the following step can detecting minutiae exactly.The work we deed in this paper has a meaning to enrich the fingerprint identification theory and to improve the performance of AFIS to a certain extent.