The Study Of Fingerprint Preprocessing Algorithms Based On Decimation-free Directional Filter Banks

Fingerprint is the most reliable research object of the biology identification techniques for its uniqueness and inflexibility. Because of its good feasibility and practicality, fingerprint identification technique has become one of the most prevalent, convenient and reliable identity authentication techniques. Fingerprint image preprocessing is the first step of automatic fingerprint identification system, whose quality will directly affect the result of automatic identification system.Using non-separable 2-D filter, the DFDFB is able to extract the edge and texture efficiently, and acquire 2~n directional information. It has widely been applied to the fields of image de-noising, segmentation, enhancement, etc. for the high directional precision and shift-invariance of its decomposition.The process of this experiment is as followed: the first, according to the characteristic of DFDFB, analyze its academic feasibility of enhancement for fingerprint images; the second, decompose fingerprint image into some directional images corresponding to different directional information based on DFDFB, and the noise in the fingerprint image is simultaneously dispersed into directional images to weaken its influence; the third, realize the foreground/background segmentation and compute the orientation field based on directional energy masking use of the local ridge characteristic, and gain favorable results compared with the methods proposed in other thesis; the fourth, process the directional images based on the directional energy and orientation field, reconstruct the processed directional images into enhanced image, and get well results through binary.The software environment is matlab7.0 and all fingerprints tested in the experiment are selected from FVC2004. Results show that the proposed method can reduce the influence of noise, link injured ridges and enhance ridges' moving shapes.