Face Recognition Using Depth Images Base Convolutional Neural Network

Abstract

Face recognition based on depth images is widely studied due to its advantages of 3 dimensional information and environment illumination insensitivity. The traditional recognition methods in this field mainly focus on hand-crafted feature design, which cannot achieve satisfactory result. In addition, there is no fixed face feature extraction method. To achieve a better face recognition performance on depth images, this paper proposes a method based on Convolutional Neural Networks(CNN). The experiment performed on database IIITD Kinect suggests that the proposed CNN architecture has better recognition performance than some traditional manual feature extraction methods, such as HOG and LBP.