Finger Vein Recognition using Gabor Features and SVM Classification

Abstract

Biometric identification using infrared vein pattern images has more distinctive features, intensity, and plagiarism issues than comparable methods. The vein pattern is authenticated by the fact that it is in a person, and cannot be seen under normal light. Internally it is immune to environmental effects and invisible under normal light, making it difficult for users to access without knowledge. A vein pattern identification using Gabor features has been implemented in this research. Specific biometric processing steps, including pre-processing, segmentation, edge detection, feature extraction, and classification of the vein patterns and current algorithms were studied. In this study, Gabor features and the SVM classification for effective recognition of vein patterns were used with 96.74 percent accuracy.