Generation of 3D facial images using video images and a facial muscle based model

摘要

A high speed, high precision method for constructing 3D human facial images Utilizing the characteristics of video images directly is proposed. A 3D facial model composed of 3D polygons and a 2D facial muscle model. is constructed from two orthogonal facial images. Video input flames are modified, and the movements of facial feature points are measured based on the contraction rule of facial muscles. The vertices of the 3D polygons are altered adaptively according to the positions of the feature points, and the modified input flame is texture mapped directly onto to the model. This method was implemented using an ordinary CCD camera and computer, and a fundamental experiment using one person's data showed its effectiveness.