Machine Learning Approach for Face Recognition from 3D Models Generated by Multiple 2D Angular Images

Abstract

We propose a machine learning approach for face recognition from 3D models generated by multiple 2D angular images that recognizes faces from multiple angle of a 3D face model. The proposed system uses SFM algorithm with SIFT detector, Approximate Nearest Neighbors (ANN) algorithm and RANSAC algorithm to reconstruct 3D from multiple RGB images. Again, it includes AdaBoost Learning algorithm that is used to train model to recognize faces and we used Local Binary Pattern Histogram (LBPH) which marks the pixels of a picture. The proposed system successfully recognizes faces with a deviation angle up to 120°, (i.e., 60° left and 60° right). Additionally, it gives an accuracy of 80% to 100% depending on angular deviation of up to from 0° to 60°. Nevertheless, the rate of accuracy of our proposed system is reversely proportional to the Angular Deviation.