Modified Binary Pattern for Finger Vein Recognition

Abstract

In this paper, the Center-Symmetric local binary pattern (CSLBP) operator is firstly used as a feature extraction method for finger vein recognition. The CSLBP feature can be viewed as a combination of the texture-based feature and the gradient-based feature. Moreover, CSLBP is easy-to-implement and computational simplicity. However, due to its small spatial support area, the bit-wise comparison therein made between two single pixel values is much affected by noise and sensitive to image translation and rotation. To address this problem, we further present a modified feature, termed Multi-scale Block Center-Symmetric local binary pattern (MB-CSLBP). Instead of individual pixel, in MB-CSLBP we perform the comparison based on average values of block sub- regions. It encodes not only microstructures but also macrostructures of image patterns, and hence provides a more complete image representation than the basic LBP and CSLBP operator. Experiments show that better performances are gained by the proposed method.