A Hierarchical Probabilistic Model for Facial Feature Detection

摘要: Facial feature detection from facial images has attracted great attention in the field of computer vision. It is a nontrivial task since appearance and shape face tend to change under different conditions. In this paper, we propose hierarchical probabilistic model that could infer true locations features given image measurements even if with significant expression pose. The implicitly captures lower level variations components using mixture model. Furthermore, higher level, it also learns joint relationship among components, expression, pose information through automatic structure learning parameter estimation Experimental results on benchmark databases demonstrate effectiveness proposed