Finger-vein identification using pattern map and principal component analysis

In this paper, we propose a new approach for finger-vein recognition which uses pattern map based on pixel-pattern-based texture feature (PPBTF) and principal component analysis (PCA). Instead of obtaining finger-vein features from multi-filtered images, we obtain the features from pattern map images. The pattern map images are generated from pattern templates which are the eigenveins obtained from PCA process. Every finger-vein image is transformed into pattern map images where edges and lines are used for characterizing the vein pattern information. PCA is then adopted to further reduce the dimension of the features and nearest neighbour is used for classification. Experiment results show that the proposed algorithm has higher identification rate compared to the existing method with only 40 features. This shows that pattern map is able to represent finger-vein pattern effectively.