mkocabas/VIBE
PyTorch implementation of a CVPR 2020 video inference model for 3D human body pose and shape estimation from monocular video.
Not currently ranked — collecting fresh signals.
star history
VIBE (Video Inference for Human Body Pose and Shape Estimation) is an official implementation of a CVPR 2020 research paper that estimates 3D human pose and SMPL body shape from video sequences. The method uses a sequence model combined with a discriminator to produce temporally consistent pose and shape estimates. It is built on PyTorch and includes training, evaluation, and demo scripts.
Frequently asked
- What is mkocabas/VIBE?
- PyTorch implementation of a CVPR 2020 video inference model for 3D human body pose and shape estimation from monocular video.
- Is VIBE open source?
- Yes — mkocabas/VIBE is an open-source project tracked on heatdrop.
- What language is VIBE written in?
- mkocabas/VIBE is primarily written in Python.
- How popular is VIBE?
- mkocabas/VIBE has 3.2k stars on GitHub.
- Where can I find VIBE?
- mkocabas/VIBE is on GitHub at https://github.com/mkocabas/VIBE.