Research On Finger Vein Recognition

Biometrics has high research value and broad application prospects in the field of authentication. Finger vein recognition, identifying people by their finger vein, is a new hotspot of biometrics. It has attracted more and more attention because of its unique advantages—non-contact, live body identification and high security.However, no biometrics is perfect, finger vein recognition also has limitations, such as, low quantity of finger vein images, especially the female’s, noise caused by illumination or finger motion during capture, poor feature information in finger vein. These problems reduce the performance of finger vein recognition, and restrict the application of finger vein recognition.In order to solve these problems and improve the accuracy of finger vein recognition, in this thesis, we proposed two solutions. The main research contents include:Many finger vein images have low quality, especially the female’s. It is difficult to improve the accuracy of these kinds of images. Considering finger vein and fingerprint both are features from finger, furthermore, they are complementary and could be captured by the same device, which is very convenient. In this thesis, we proposed a fusion method which fuses finger vein and fingerprint for recognition in order to improve the accuracy of authentication system. Comparing the levels of fusion, score level is chosen to fuse fingerprint and finger vein recognition. Minutiae based methods are adopted in fingerprint and finger vein recognition, and the match scores are normalized before fusion. Experimental results show that the fusion of fingerprint and finger vein leads to a dramatically improvement in performance.Considering the problem of poor feature information and segmentation errors, we choose binary pattern to improve the performance of finger vein recognition. A new binary pattern method, finger vein recognition based on personalized best bit map (PBBM), is proposed. Different from other traditional binary pattern, which has the same weight of bit, PBBM uses consistent bits for personalized match, i.e., the weight of the bit in codes are different, which can overcome the problem of noise and bit waste. Two new conceptions were defined in the thesis, one is "the best bit", and the other is "the best bit map", and the personalized best bit map (PBBM) was generated from them. However, as research on binary pattern finger vein recognition is fairly new, and the fusion of finger vein and other recognition is in the immature stage, there is still much work to do. Future work involves exploring fusion of finger vein recognition which is based on PBBM or other binary pattern with other biometrics, and utilizing authentication system to have high recognition rate and high security.