Abstract:

A new method for dorsal hand vein recognition based on FRAT was proposed. We firstly partitioned these hand vein images into blocks on the basis of  their “line” features; then did FRAT on the block images, and put up a new method to get the maximum by means of the  nonequilibrium distribution of vector in order to get the texture features; finally,  obtained classification results through feature matching. Compared with discrete Radon transform and traditional Hough transform, our method improves the detection speed of straight line features. At the same time, our method was applied to the test of  800 samples, showing a recognition rate of  96.5%. It has demonstrated that this algorithm is effective.

Key words: finite Radon transform (FRAT), dorsal hand vein, recognition algorithm