The paper deals with an algorithm of texturing of the surface of 3D human head model. The proposed algorithm generates a texture consequently of several camera frames of the input video sequence. The texture values from the camera frames are mapped on the surface of the 3D human head model using perspective projection, scan line and 3D motion estimation. To decrease the number of camera frames a filling of empty places by a simple interpolation method has been done in the texture plane.
展开▼