Sketch Map
The Pipeline of Our Proposed Auto-encoder.
Side-by-side Comparisons of Image-based Facial Reconstruction Results.
Abstract: We develop a new algorithm to perform facial reconstruction from a given skull. The target application is the forensic body identification from skeletal remains. We reconstruct 3D face candidate models from a portrait database, then do a superimposition to get a best-matched face, and finally revise this face according to the superimposition. To support this face revision, we build a generative model for constrained 3D face inpainting. Our experiments demonstrated that the proposed pipeline is more stable and accurate than existing approaches.
Upload your file(*.jpg)
The Pipeline of Our Proposed Auto-encoder.
Side-by-side Comparisons of Image-based Facial Reconstruction Results.
These are just demo colors. You can easily create your own color scheme.