A Fingerprint Matching Algorithm Based On Reference Point Of The Center Area And Implemention

With the development of information technology in recent years, Biometrics identification technology has been drawn wide attention. Compared to other Biometrics, automatic fingerprint identification technology performs better in accuracy, utility and security, so it is widely used in various commercial products.Although the research on automatic fingerprint identification technology has been engaged for many years, the technology has a long way to be satisfied due to defects, noise and non-linear deformations, the algorithm will have to be improved. Based on the existing data this paper makes a deep research on several key algorithms.1. The orientation field lower the dimension of the 2-D Gabor filter, making enhancement processing with only two one-dimensional Gaussian filters by convolution and making the calculation more easy. The experiments improve that the enhancement algorithm can not only resist the noisy but also retain the fingerprint ridge and valley structure totally.2. With the use of the original distribution of pixels parallel or perpendicular to the orientation field, the binarization is presented self-adaptively. The experiment proves it is superior to the block-Ostu binarization.3. Compared with rapid thinning algorithm, this paper proposes a thinning algorithm by looking up table, which can make the ridge to be single-pixel-wide. After that, the ridge structure which may give rise to false minutia can be eliminated by ridge tracking and troubleshooting. And the number of false minutia could be reduced largely.4. This paper proposes a fingerprint matching algorithm based on reference point of the central area. It gets the reference point by scanning the end point or bifurcation point in the center area, and the parameters of rotation and translation could be calculated easily. So the two-point feature set could be put under the same coordinate system for matching by a certain deviation tolerance. The Experiment proves the feasibility of the algorithm, while the algorithm also achives good performances in the test of fingerprint library.5. Experiments have been done in the PC of software VC++ 6.0 and the results prove that the identification algorithm is available and can achieve the desired results.6. Finally, summarize the work of the thesis, analyze the improvements need to be done and give the direction of future work.