Palm Vein Recognition Using Enhanced Symmetry Local Binary Pattern and SIFT Features

Recently, palm vein recognition is new biometric technology with a high degree of privacy and security because this technique uses the blood vessels under the palm skin to establish identification. This paper proposes a novel palm vein feature extraction method for contactless palm vein recognition based on combining enhanced center-symmetric local binary pattern (ECS-LBP) with SIFT, called EL-SIFT. The proposed method includes two steps: (1) applying ECS-LBP to detect stable and clear palm-vein lines; and (2) extracting SIFT feature on palm vein lines image. The experimental results on the public contactless palm vein databases (CASIA Multi-spectral Palm vein Image Database V1.0) show that our proposed method is accurate and robust for palm vein recognition in comparing with other approaches in the literature.