Studies On Automatic Fingerprint Identification Based Texture Feature

Fingerprint (FP) recognition is one kind of biometrics. With a FP sensor capturing image and FP compared on a computer, identity can be accomplished. It is a hot spot and the most widely used branch of biometric with many advantages, such as easier to use, higher reliability and lower cost. Automatic Fingerprint Identifications System (AFIS) has so bright future on ID authentication with the development. This paper use Gabor filter to extract the texture feature of FP and match. FP image is a special image that has characters of direction and frequency. Fortunately, Gabor filter also has characters of direction and frequency selectivity. We use the local direction and frequency of the FP image as the Gabor filter function's parameters, compute the convolution of the Gabor filter and the FP image to extract texture feature from the center area of the FP image, and then use the araneose grids carve up these texture feature image, the texture feature info in all of these grids makeup a compact fixed length finger code. Feature matching is just calculating distance between the corresponding finger code. In this method, the feature elements capture the local information and the ordered enumeration of the grids captures the invariant global relationships among the local patterns. Avoid some faults of the Minutiae-based matching method: So many preprocessing and every step affect the final result of system. Feature vector don't has fixed length and can't be stored, indexed. Minutiae feature is the unsteadiness info. Experimental results show that Structure-based FP recognition method not only makes best use of texture structure information contained in the gray level FP image but also reflect the local information. And this method rely fewer on image quality than the other, almost do not need reprocessing, it is proved to be relatively high effective in recognizing damaged or blurry fingerprint images, and feature vector has fixed length, so it make storing, indexing and matching easy. In addition this method have the effect that overcome the image circumrotate and move affection in a certain extent. The system is relatively high valuable in practice.