Image Processing for Biometric Scanning of the Palm Vein Pattern

Abstract

Due to increased requirements for access control systems, the use of biometric recognition technologies is becoming a reliable solution for the protection of critical information. One of the best ways of personal identification is to use the palm vein structure. The chapter deals with improving accuracy in the problem of recognizing the palm vein pattern when comparing biometric templates using the Canny edge detection algorithm and the Gabor filter. 2D Gabor filter improves the adaptability of recognition and is therefore proposed to solve the problem of image blurring and select a threshold when the traditional Canny algorithm smoothes the edges. The results of experiments show that this filter can detect less pronounced edges and provides more complete information about the image, which has a positive effect on the result of biometric authentication. The similarity of two biometric templates is determined using the Minkowski metric. Experiments conducted on the original facility show high performance, as well as good results in false acceptance errors (FAR = 0%) and false rejection errors (FRR = 0.01%) based on processing 360 images captured from 26 people, which makes it possible to use the proposed method in the identification and authentication system at existing data security facilities.