Yuming Gu(顾育铭)

This project implements a way to reconstruct personal identity head back from light stage data. We first use silhouette images to get a visualhull model. Then based on this model we use CGAL library to get the intersects between the light-stage model and the visualhull model and finally get the correspondence and reconstruct identity-preserved headback .

 This project we first find the similar face model from a light-stage database, then we rendered it using rendering engine. After that, we use AI-replacement tech to swapping the face by using a sequence of the target face to train a model. Then we use traditional method(color transferring, boxing filter) to enhance the quality.