A study of real-time image processing method for treating human emotion by facial expression

摘要

The aim of this study is to develop image processing methodologies for recognition of human facial expression to understand human emotion, and for the synthesis of facial expression to express emotion by computer graphics. With respect to the recognition of facial expression, a real-time image processing system is developed to calculate the distance changes of various feature points on the face and then correlate them with fuzzy rules to infer the degree of six basic emotion factors. On the other hand, a so-called muscle model process is adopted for effective image synthesis for facial expression, muscle contractions are simulated and animated as facial expression, based on a facial action coding system.