Frontal Face Synthesis Based on Multiple Pose-Variant Images for Face Recognition

Abstract

Pose variance remains a challenging problem for face recognition. In this paper, a stereoscopic synthesis method for generating a frontal face image is proposed to improve the performance of automatic face recognition system. Through this method, a frontal face image is generated based on two pose-variant face images. Before the synthesis, face pose estimation, feature point extraction and alignment are executed on the two non-frontal images. Benefited from the high accuracy of pose estimation and alignment, the composed frontal face retains the most important features of the two corresponding non-frontal face images. Experiment results show that using the synthetic frontal image achieves a better recognition rate than using the non-frontal ones.