Finger vein recognition using local directional pattern

This paper presents finger vein recognition using different types of local directional patterns to gain best result in terms of classification accuracy, Equal Error Rate and to remove the limitation of the Local binary pattern which is its sensitivity to noise. In this paper we analyzed different techniques of LDP to make finger vein recognition reliable. LDP is used to compute the edge response in all eight directions and considering each pixel position. The performance is evaluated on the SDMULA database and data set of 3816 samples of finger vein images. After all the computation and process, we analyzed the best result with an EER to 14.3082% and classification accuracy of 86.1635%.