The Research Of Fingerprint Image Singular Points Detection And Minutiae Extraction Algorithm

Among many biometric technologies, fingerprint identification technology is one of the earliest developed and the most widely used technologies. So it is the significant part of individual authentication in the future. Automatic fingerprint identification technology mainly related to image capture, fingerprint image processing, feature extraction, feature classification and matching process. Fingerprint classification and matching are based on the fingerprint feature points achieved. The former is based on singular points of fingerprint characteristics, while the latter is based on fingerprint minutiae. Therefore, fingerprint feature extraction is a very crucial process in fingerprint identification technology.Although algorithm of fingerprint recognition have larger development at home and abroad, people requirements of the automatic fingerprint recognition technology are rising steadily as increasing demand. So studies on the theory and application of fingerprint recognition have significant meanings both in theory and application in different needs. In this paper, summing up the results of existing research in this field, two key links as the fingerprint singular points detection and the minutiae feature extraction are in-depth researched.First of all, the paper describs the existing singularity detection algorithm, introduces the classic detection algorithm based on the Poincare index, studies the detection algorithm of which is based on the abundance direction.By the simulation Matlab to compare the results, based on the richness direction of algorithm has obvious advantages in detecting singular points correctly and the algorithm computing time, the performance of the entire identification system has some improvements.Secondly, during the fingerprint details extraction, the paper studies two kinds of algorithms respectively, which are more advanced the algorithm based on ridge track and based on 8-neighborhood codes ridge track. Experimental results show that the extraction of the latter has more minutiae and higher accuracy, the extraction of the former is higher speed. In practice, people should select the compatiable extraction algorithm according to the overall demand of recognition system.Finally, based on this problem,which extracted the endpoint and bifurcation points contain a number of pseudo-feature points, a method aiming at different types of pseudo-feature points to remove respectively is proposed. Through analyzing the characteristics of pseudo-feature points,it gives the effective maps of removing pseudo-feature points and establishs a good foundation for follow-up fingerprint match.