Performance analysis on feature extraction using dorsal hand vein image

Abstract

Personal verification has become a severe and necessary perspective of security access systems. Compared to the traditional authentication codes, the personal identification using biometric feature has attracted more attention these years because the biometric feature is hard to forget and filch and easy to use. Biometrics refers to the identification of the humans by their characteristics or traits. This paper work considers vessels at the back side of the hand called dorsal vein. The features like local binary pattern (LBP), histogram of oriented gradients (HOG) and Weber local descriptor (WLD) are extracted, and performance analysis is done in terms of K-NN classification accuracy. The 10 image of right dorsal hand vein and 10 image of left dorsal hand vein were used from 102 persons, including both male and female. There were 50 males and 52 females. The total number of images is 2040. The WLD method has recognition accuracy up to 98%, the LBP method has 96% of recognition accuracy, and compared with both, the HOG method proclaims the highest recognition accuracy up to 99%.