Robust pose estimation of face using genetic algorithm

Abstract

This paper focuses on the problem of human face pose estimation using single image. Because the traditional approaches for 2D-3D feature-based pose estimation problem requires two inputs, they can not work very well due to lack of correspondences of the input image. In this paper, we propose pose estimation algorithm for human face based on genetic algorithm. The proposed method overcomes the shortcomings by using a general 3D point-feature template as the correspondences and some more constraints onto the optimizing function. The experiments show the proposed method gives good performance in accuracy and robustness. © Springer-Verlag Berlin Heidelberg 2004.