부산대학교 도서관

In the identification system of finger vein, the extraction of the region of interest (ROI) is a very important step. It determines whether the ROI of the vein image of the same finger has a high similarity, and then directly affects the following two steps of feature extraction and recognition and matching. This paper used deep neural network for the ROI extraction of the finger vein, and proposed a ROI extraction algorithm for finger vein based on capsule neural network aiming at the problems of information loss in traditional convolution neural network. By using two kinds of public data sets for the test, it can be seen that the algorithm of this paper has a good effect on ROI extraction of finger vein images.