Contact-Less Palm Vein Recognition Based on Wavelet Decomposition and Partial Least Square

To solve the problem of low recognition performance caused by contact-less imaging and poor palm vein image quality, a novel recognition method is proposed. Firstly, the ROI based on the thenar (a part of palm) is located; secondly, the palm vein features based on wavelet decomposition and partial least square are extracted; finally, the images are matched by Euclidean distance. In self-build palm vein database, the experimental result shows that the best recognition rate of this method reaches 99.86%. Comparing with the other typical palm vein recognition methods, the performance of proposed approach is the best. In conclusion, the scheme can improve the identification performance of contact-less palm vein recognition significantly.