Abstract: Aiming at the problem that the traditional LPQ( Local Phase Quantization) feature extraction algorithm can not extract the details of the hand vein image，the region is divided into sub-blocks for LPQ feature extraction according to the characteristics of the vein texture image． Firstly，the back vein texture image is divided into 9 equal-sized sub-images． Then，the LPQ feature extraction algorithm is used to extract the features from the dorsal vein of the opponent，and the extracted vein texture information extracted from each sub-region is integrated to form a whole vein image． The vector feature is then classified using the nearest neighbor classifier to the samples in the dataset． The experimental results show that the highest recognition rate is 96． 50% when the number of blocks is 4 × 4．

Key words: hand vein recognition, feature extraction, phase quantization, image processing