Study On The Key Algorithms In Automatic Fingerprint Identification

With the rapid development of intelligence and automation, the traditional identification technology can not meet the demand for speed and accuracy more and more, on the contrary, modern biological recognition technology based on the information processing becomes a research hotspot. The automatic fingerprint identification has become one of the most extensive applications of biometrics because of its superior properties. Although the research on automatic fingerprint identification technology has made some progress, how to improve the accuracy and speed is still an urgent problem due to the rapid expansion of information.Automatic fingerprint identification is to determine the identification of the fingerprint by automatic processing the relevant feature information in fingerprint image through integrated image processing and pattern recognition, computational intelligence and other fields of knowledge. Based on the analyzing and summarizing the results of previous studies, the fingerprint image enhancement and feature matching method have been studied deeply in this paper. The automatic fingerprint identification system (AFIS) has also been designed and implemented, which integrates all of the research.(1) In order to keep more texture features of fingerprint while denoising, a new fingerprint enhancement method based on wavelet transform and unsharp masking was proposed in this paper according to the multiresolution and local time-frequency analysis of wavelet. First of all, a fingerprint image was decomposed by wavelet, and then some different treatments were made by the characteristics of the different frequency graphs. Finally, the initial enhanced fingerprint image was reconstructed by the wavelet coefficients which had been adjusted before. In addition, for highlighting the details, unsharp masking was used to process the fingerprint image. Experimental results show that this method can filter out the noise while enhance the texture features and thus improve the quality of the images.(2) Because the location of the fiducial point is inaccuracy and time-consuming in fingerprint matching, a new matching algorithm based on improved fiducial point locating is proposed in this paper. The algorithm constructs a local structure with the help of the core point, and figures out a fiducial point in this structure by using the principle of congruent triangles. And then all the minutiae features are expressed in polar coordinates, at last the final decision can be made with a changeable boundary box method. Experiments prove that the location of the fiducial point is more accuracy and the time is saved, thereby the recognition rate and efficiency are enhanced(3) Based on detailed requirements analysis, Automated Fingerprint Identification System has been designed and implemented, which integrating all the results in this paper. The system can visually display the user's operating results, more convenient and practical, and create a foundation for the future practical application.