A Face Recognition System Using Support Vector Machines and Elastic Graph Matching

Abstract

An efficient face recognition system by the combination of Support Vector Machines (SVMs) and elastic graph matching was presented. The implementations of this system are as follows: Firstly, the two eyes of a face image are detected by using SVMs approach, and the detected eye coordinates are used as reference points for alignment and normalization of the face image. Secondly, elastic graph matching is performed to locate the facial feature points. The preprocessed face images have the same size and the eyes lie at the uniform positions, elastic graph matching is limited to local distortions only, so the matching time can be reduced considerably. Lastly, local facial features are extracted for face register or recognition, recognition is performed by comparing the similarity of the local facial features of a test image to all trained images. The experimental results demonstrate the effectiveness of this face recognition system.