3D facial reconstruction system. Using a video and a prior collection of photos of the persons face from various angles, this system can construct a 3D model of the persons face as they are speaking. The system works by using the 2D photos to form a 3D base shape construction of the face, then doing pose estimation to map the 3D base shape to the frames of the video, then it has to deform the 3D base shape for the face movement, and finally there is a shading refinement step to get the final image to look right. The system can be used to create video of the person as if they were filmed from a different angle, but it apparently cant map one persons facial movements to a different persons face, which would make it possible to make video of a person saying something they never said.
Posted on: Thu, 25 Sep 2014 00:25:40 +0000
Trending Topics
Recently Viewed Topics
© 2015