GitHub - Lotayou/densebody_pytorch: PyTorch implementation of CloudWalk'...
source link: https://github.com/Lotayou/densebody_pytorch
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
README.md
densebody_pytorch
PyTorch implementation of CloudWalk's recent paper DenseBody
Prerequisites
Ubuntu 18.04
CUDA 9.0
Python 3.6
PyTorch 1.0.0
chumpy (For converting SMPL model to basic numpy arrays)
spacepy, h5py (For processing Human36m cdf annotations)
(Optional) Install torch-batched-svd for speedup (Only tested under Ubuntu system).
TODO List
- Creating ground truth UV position maps for Human36m dataset.
- Finish baseline model training
- Testing with several new loss functions.
- Testing with different networks.
- Report 3D reconstruction results.
- Setup evaluation protocal and MPJPE-PA metrics.
Current Progress
Finish UV texture map processing. Here's the result:
Align SMPL mesh with input image. Here's some results:
Citation
Please consider citing the following paper if you find this project useful. DenseBody: Directly Regressing Dense 3D Human Pose and Shape From a Single Color Image
Disclaimer
Please note that this is an unofficial implementation free for non-commercial usage only. For commercial cooperation please contact the original authors.
Recommend
About Joyk
Aggregate valuable and interesting links.
Joyk means Joy of geeK