An Research And Implementation Of Fingerprint Recognition Based On Texture Analysis

With the continuous development of computer and information technology, biometric identification technology has been widespread concern. Fingerprints have become one of the most reliable study subjects of biometric identification technology because of their uniqueness and invariance for life. There are many final products on the heel of this technology now, but because many core technologies could not be released due to commercial interests, and social development has put higher requirements on fingerprint identification, the study on the fingerprint recognition technology still has higher realistic and theoretical significance.This paper has systematically studied fingerprint identification theory and technology based on texture analysis; studied and made a realization of a Gabor filter texture feature extraction and a matching fingerprint recognition algorithm. The algorithm first inputs fingerprint image and then has a gray-scale regulation and a calculation of the direction of the market, furthermore, it locates the reference point on the basis of the center of the fingerprint image, and then cut the fingerprint image to extract effective region centered on the central reference point, making follow-up processing focus on effective areas. According to the characteristics of fingerprint images, fingerprint images can be seen as a direction of texture. We can make use of Gabor filters'features of the choice of direction and frequency to regard the partial texture feature map and partial frequency as parameters of Gabor filter function. First, we can transform fingerprint images into a multi-directional Gabor filter to get a partial texture feature map of the fingerprint image. and then we encode fingerprint characteristics on the texture feature map, so that the fingerprint matching can be simplified into Euler distances between the calculating characteristic of inter-coding distances. The method makes the most of the partial fingerprint image features and global features to solve poor-quality fingerprint identification problems and to overcome many shortcomings based on the characteristics of minutiae-based matching algorithm, and the method is simpler and easier to implement.This paper has designed a fingerprint identification system, using Matlab7.5 to achieve computer simulation. The experiments indicate that the algorithm in this article not only makes best of rich texture information in gray-scale images, but also reflects the partial fingerprint information. And this method is less dependent on the quality of the image, and overcome the impact of the fingerprint translation and rotation on the fingerprint matching in a certain range. The algorithm has high practical value.