Most of the facial animation applications, such as automatic face detection, recognition, and tracking are sensitive to the different lighting conditions and complicated background environment, especially for accurate expression analysis in virtual conferencing system. The main work in this research is to detect the faces under these complex situations and reduce the influence of different lighting. In addition, the pose estimation is the bottleneck of the whole framework, the speed of which is to be improved. An adaptive skin color model, scanning over the downsampled skin map, a modified mean shift algorithm and a new skin mask are used to refine and speed up the head tracking procedure. The lighting distribution is estimated by a second order function, and the lighting effect is compensated. Finally, we modify the good feature selection algorithm to the texture map to pick out the features, and then perform the analysis-by-synthesis pose estimation with the texture map of good features to speed up without losing accuracy
展开▼