A Palmprint Recognition Algorithm Based on Harris Synthetically Method

Abstract

The algorithm gets corner sets from palmprint images using a customized Harris method. Then apply distance and orientation features to interconnect the corner points with lines and optimize the result. Fuzzy blocks are used to increase the match probability in the recognition algorithm. The algorithm chooses different filter windows to influence the result corner sets. Then integrating with distance and direction features recognizes principal palmlines. This paper takes tree structure as storage organization. Palmprint image requires about 250 bytes and saves more than 50 percent storage space compared with other algorithms. Experiments show that the algorithm is complete easily and having a good recognition with a decrease on the amount of memory space.