Palm Vein Detection Based on Local Gabor Vein Code Pattern and Hog

Abstract

Nowadays, Palm vein biometric is one of the emerging techniques to build up a well-organized recognition system. This paper focuses on quality-based palm vein detection. To recognize hand palm vein images, Local Gabor Vein Code Pattern (LGVCP) is proposed. It comprises of three modules, namely Preprocessing, LGVCP and Palm Vein matching. Preprocessing is performed using Otsu thresholding method and Contrast Limited Adaptive Histogram Equalization (CLAHE) techniques. Palm vein features are extracted by applying LGVCP and Histogram of Oriented Gradients (HOG). Chi-square distance and Euclidian distance values are applied for palm vein matching. Experiments are performed on CASIA Palm vein Image database. The performance of the proposed method is evaluated by metrics, namely accuracy, specificity, sensitivity, error rate, recall and precision. The proposed method provides an accuracy of 99.001% compared to the existing ones.