Research And Application Of Fingerprint Image Enhancement And Matching Algorithm

With the city expanding and increasingly well-developed transportation network in modern society, urban population is increasing rapidly and showing more and more mobility. How to identify ID simply and effectively for such a large-scale population is an important issue that government must take into account. In addition, with the high development of computer and network technology, information security has shown the unprecedented importance. Biometric identification is necessary to ensure information system security. Finance, national security, justice, e-commerce applications will require accurate identification and authentication. Automated Fingerprint Identification System(AFIS) is drawing more and more attention for its easy-to-use, friendly interface and low prices. There were a lot of fingerprint recognition products, but most of them could lose efficacy considering the balance of accuracy and real-time requirement. This paper mainly described the Automated Fingerprint Identification System around two key technologies: fingerprint enhancement and matching. According to the shortcoming of the past algorithms, the paper improved algorithms of the fingerprint enhancement and matching. After performance analysis and comparison, the new algorithm showed better ability.This paper introduced the history of fingerprint recognition technology, the status quo and development trend. After analyzing the advancement and shortcomings of the previous enhancement algorithms, the adaptive fingerprint enhancement algorithm based on frequency domain and the hybrid filter fingerprint enhancement method were presented. Two algorithms were all based on the fingerprint image characteristics in the frequency domain and spatial domain. The filter was designed according to the turn of the fingerprint ridge pattern. Experiment shows that the algorithms are better than past algorithms. A fingerprint recognition algorithm based on hybrid ant colony was presented. When the fingerprint had a small deformation, fingerprint matching algorithm only confirmed translation, rotation and scale parameters. Ant colony algorithm was a very good optimization algorithm, but the basic ant colony algorithm had the problems of long-time consumption and easily getting into the local optimum. So the simulated annealing algorithm was used to accelerate the optimization of the fingerprint. The annealing and backfire procedure could increase the diversity of solutions to avoid a local optimal solution. A novel algorithm for fingerprint recognition based on RANSAC strategy was presented. The local characteristic of the fingerprint was used to find the coarse matching points. Then the RANSAC algorithm was used to match the fingerprints. In the process of fingerprint matching, compensation method is used to eliminate the impact non-linear deformation.