Research On The Algorithm Of Fingerprint Recognition System

The authentication of fingerprint recognition is a personal recognition technology that undergoes historical test, the mankind uses fingerprint to carry on the identity recognition for a history of several hundred years. Nowadays, Fingerprint Recognition System based on its reliable, unique biological characteristic, regarded information treatment technology as the means, combine biotechnology and information technology organically together, make it have more and more extensive application in people's life. Though theoretical research and application development of Fingerprint Recognition System have made great progress, but the recognition accuracy and processing speed of the algorithm still remain to improve. Constructing the reliable, security and practical Automatic Fingerprint Recognition System accurately becomes the focus of research.Based on the analysis of domestic and international research results in recent years about Automatic Fingerprint Recognition System, this paper makes a research on the algorithm of Automatic Fingerprint Recognition System, mainly contains three major parts: fingerprint image preprocessing, fingerprint feature extraction and fingerprint matching.The present study mainly explores the improvement of recognition rate and processing speed of Fingerprint Recognition Algorithm. Based on the analysis of traditional means of recognition, it makes researches on the enhancing of the fingerprint lines and the template matching of features. At first, due to the fact that the enhancing algorithm is hard to be satisfied both in effect and efficiency, the study proposes a algorithm for the enhancing of the fingerprint lines which is good at effect and efficiency. Algorithm uses Gabor filtering function to enhance the fingerprint lines, which utilizes the orientation and frequency characteristic of the lines; Algorithm dramatically improves the enhancing effect and decreases time-consuming by the means of changing the operation of two-dimentional convolution into two times of the operation of one-dimentional convolution, making the continuous orientation into 16 discrete direction; Besides, against the feature matching, the present study put forward a matching method based on the line regulation, which puts features into polar coordinate for matching; For the choice of a pair of reference points, the study has a different treatment for the situation of having a singular point and non-singular point; Then through the judgement of similarity of lines to confirm whether the points is corresponding or not; After adjustment of the polar coordinate, it makes an adaptive threshold method by designing a variable border box for matching, then a Match_score will be got, compare Match_score with a Threshold to determine the matching result.All the algorithms have been tested in FVC2000 standard fingerprint database. The experimental results indicate that the Fingerprint Recognition System has been improved in the study, the processing speed and recognition rate have increased significantly, and the system has a very good effect on robustness.