"real-time approach for mapping all human pixels of 2D RGB images to a 3D surface-based model of the body." "Our work shows that one can efficiently compute de

"real-time approach for mapping all human pixels of 2D RGB images to a 3D surface-based model of the body." "Our work shows that one can efficiently compute dense correspondences between 2D RGB images and 3D surface models for the human body. Unlike common works in human pose estimation that operate with 10 or 20 human joints (wrists, elbows, etc), this work accounts for the entirety of the human body, defined in terms more than 5000 nodes. The resulting speed and accuracy of our system accelerates connections with augmented and virtual reality. Earlier works on this problem would require computation in the order of minutes, initialization by an external system e.g. for human joint localization, while being particularly brittle. DensePose operates at multiple frames per second on a single GPU and can handle tens or even hundreds of humans simultaneously." GitHub: https://github.com/facebookresearch/DensePose https://github.com/facebookresearch/DensePose