Multi-scale Bayesian face recognition by using anti-symmetrical biorthogonal wavelets
Jinye Peng, YU Bian-zhang · 2002
In this work, we analyze theoretically the derivative feature of ASBW by which the gradient vector of a facial image at every resolution level can be obtained. Then an analysis of the robustness of directional gradient angles of facial images is presented. Then we propose a multi-scale Bayesian face recognition based on ASBW. The effectiveness of our method is proved by experimental results.