3D Facial Expression Classification Based on Self-Organizing Mapping Network

In this paper, we propose and explore a novel method to recognize human facial expression in 3D. Based on a shape descriptor to express shape in 3D converted from the original 3D points cloud, we localize the different features in 3D shape around the nose tip. Then we exploit Self-Organizing Map(SOM) to recognize the similarity of the same expression and the variations between different expresses. Experiments performed on data with facial expression variations show that our method is able to separate different expressions.