In this paper, we try to make automatic interpretation of facial expressions on grey level images. The first step is a semi-automatic extraction of some facial parameters like the size of the mouth (both width and heigh) or the presence wrinkles (on the nose, eye corners, front ...). All these parameters are aggregated using fuzzy techniques to give a final curve that characterises the expression. It is then possible to answer simple questions like "is it a smiling face, an angry face ... ?".
展开▼
