抄録

H-047
A Study of Extracting 3D Facial Feature from Kinect's Image by Integrating ASM and Depth Map

The paper describes an approach to 3D facial feature extraction using ASM and depth map. We use color image and depth map which are captured by Microsoft Kinect. Firstly, standard ASM is applied in color image to find 2D facial feature. Then we obtain and modify the feature points’ depth information from calibrated Kinect’s depth map. We experiment our proposal through integrating 2D facial feature points with depth information. And the future work of the study is to warp facial textures onto 3D feature points by Kinect’s image in order to reconstruction 3D human face.