Recognition of Human Facial Expression Based on 3D Data
Wang Zeng-fu · Jisuanji fangzhen · 2005
To make full use of the 3D image, based on the former research works, this paper proposes a new 3D-data-based method for recognizing human facial expressions. In this method, according to the characteristics of 3D image instrument, 3D data are used directly to identify feature points, and traditional procedure of transferring 2D image to 3D image is avoided. To extract the feature points, a new matching method using templates are found based on the characteristics of 3D image. Combined with original 2D images, expressions are recognized by analyzing the characteristics of feature points. By this method, six facial expressions can be recognized in a high rate in the experiments.