Fingerprint Recognition System based on Orientation and Texture Features

Fingerprint recognition is one of the mostwell-known and publicized biometrics forpersonal identification. Fingerprints exhibitoriented texture-like patterns. The textureinformation of the fingerprint can be used forfingerprint matching. Gabor filters can optimallycapture global and local texture informationeven from poor-quality or incomplete images.But Gabor filterbank-based approach use onlytexture information for fingerprint recognitionand it is not robust to image distortion androtation. In this paper, a hybrid fingerprintmatching algorithm is developed for identifyingthe low quality fingerprint images by combiningorientation features and the local texture patternobtained using a bank of Gabor filters. Theproposed matching approach is compared withthe filterbank-based approach, and the proposedsystem produces a much improved matchingperformance by combining the orientationfeatures to the filterbank-based features