Research of Fingerprint Preprocessing and Recognition Algorithm
|Course||Applied Computer Technology|
|Keywords||Fingerprint Image preprocessing Image Segmentation Gabor filters Triangle topology|
Now, with the development of e-commerce, the traditional password-based security systems or keepsake can no longer meet the needs of society and therefore a need for a safer, stronger anti-fake security systems. Biometric because of its unique and permanent, very suitable as a basis for authentication. In many biometric inside, because the fingerprint has its unique advantages, fingerprint recognition technology has become the direction of computer pattern recognition a very hot area of ??research, market prospects are very broad. Building a fingerprint pattern recognition system, the fingerprint image preprocessing is the first step, the quality of pre-recognition systems have a direct impact. This paper studies the fingerprint image preprocessing in image segmentation, image enhancement and image thinning parts. In the segmentation part, the collected image by calculating the intensity field for image segmentation; enhancement and refinement in part in the study and research scholars at home and abroad are summarized based on the results, to achieve a variety of enhancements and refinement algorithm to analyze the various advantages and disadvantages of two algorithms. While the part of the algorithm has been improved, improved operational efficiency and processing algorithms have varying degrees of quality improvement. Improved algorithm is as follows: (1) in the fingerprint image enhancement algorithms, from two aspects of the enhancement of existing algorithms: spatial enhancement algorithms and frequency domain enhancement algorithms. Were achieved based on median filter enhancement algorithm and STFT enhancement algorithms. Gabor filters based on the analysis of fingerprint image enhancement algorithm, the right to be improved algorithm is proposed an improved wavelet-based Gabor function enhancement algorithm, the two-dimensional Gabor function into two one-dimensional function, reducing the computational , improving the speed of the algorithm. (2) in the fingerprint image thinning, the introduction and implementation of common OPTA thinning algorithm (including the classic OPTA OPTA algorithm and the improved algorithm) and fast thinning algorithm. Analysis of the bifurcation point OPTA thinning algorithm is not complete and glitches too fast thinning algorithm produces refined and not thorough enough problem, we propose a pattern-based fingerprint thinning algorithm, experimental comparison shows that the algorithm speed and burrs and other aspects of an improvement over the previous two algorithms. Finally, to illustrate the entire fingerprint pattern recognition system preconditioning effect on subsequent links, describes some of the classic fingerprint feature extraction algorithm and triangle-based topology matching algorithm. In this paper, Visual C 6.0 build of the whole fingerprint pattern recognition system, the various algorithms also been achieved on this platform, better, for the follow-up study has laid a solid foundation.