Finger vein recognition based on zone-based minutia matching

Abstract

Finger vein is universally regarded as a secure and convenient biometric pattern. Consequently, it has received considerable attention from researchers and is widely used in real world applications. However, the undesirable image quality and deformation problem have limited the improvement of its recognition accuracy. In the literature, minutia-based finger vein recognition is generally acknowledged as deformation-tolerant, but problems still remain: 1) the minutiae extracted in some finger vein images are small in amount; 2) one minutia is compared with all the minutiae from other images during matching, which is time-consuming and prones to false pairings. In this paper, a zone-based minutia matching technique, which combines minutia matching with traditional region-of-interest (ROI) based method, is designed to deal with these problems. First, we extract minutiae from each segmented image block to ensure the amount of the extracted minutiae. Second, minutiae in a rational neighborhood zone are selected for matching, which discards unnecessary matchings and avoids false pairings to some extent. Compared to traditional methods, the designed matching procedure is also parameter-free and more stable when conducting minutia matching. Extensive experiments demonstrate the robustness and efficiency of the proposed method. The results also show the superiority of our method over other similar works.