Finger Vein Recognition based on Histogram of Oriented Gradients (HOG)

Abstract

In this paper, a Region of Interest (ROI) extraction method is proposed based on labeling vein images using morphological processing. Firstly, finger vein images are segmented to remove the unwanted background or the shape of the device. Secondly, the images are oriented to correct to solve the finger displacement’s problem. Thirdly, ROI localization method is used to accurately extract the region of vein vessels. Finally, Histogram of Oriented Gradient (HOG) features are extracted to recognize that person is the genuine or imposter. Segmented finger vein and calculated orientation can support each other to produce higher accuracy in localizing ROIs. In addition, a simple feature differencing method is proposed to reduce the calculation time for matching.