Finger Vein Recognition Using Pattern Matching and Corner Detection Strategies

Now-a-days, storing places are given with some kind of protection for security purpose like password, pin number or using any biometric identification system. For personal authentication, identification systems are used which utilizes Finger vein physiological biometric technology. This type of authentication is based on finger vein pattern’s physical feature. In conventional techniques, combination of genetic algorithm and selection based on correlation filter are used to generate user-specific threshold in branch tracking step. Improved fuzzy clustering algorithm is used for deciding nearest points between samples. However in the vein extraction stage exact finding of the corner becomes very difficult task, to solve this issue, corner detection algorithm is utilized for features extraction (corner points) from the images of finger vein, where pattern matching is based on corner difference which is represented in point form using neural network classifier. Analysis is carried out on database of Hong Kong Polytechnic University (HKPU) to demonstrate the robustness of proposed technique with respect to accuracy, specificity, sensitivity, recall and Precision.