#Face-reconstruction
Showing 5 of 5 repositories tagged #face-reconstruction, ranked by stars
Learning to Regress 3D Face Shape and Expression from an Image without 3D Supervision
Official repository accompanying a CVPR 2022 paper EMOCA: Emotion Driven Monocular Face Capture And Animation. EMOCA takes a single image of a face as input and produces a 3D reconstruction. EMOCA sets the new standard on reconstructing highly emotional images in-the-wild
A high-fidelity 3D face reconstruction library from monocular RGB image(s)
This is a implementation of the 3D FLAME model in PyTorch
Tensorflow framework for the FLAME 3D head model. The code demonstrates how to sample 3D heads from the model, fit the model to 2D or 3D keypoints, and how to generate textured head meshes from Images.