We provide the config files for VIBE: VIBE: Video Inference for Human Body Pose and Shape Estimation.
@inproceedings{VIBE,
author = {Muhammed Kocabas and
Nikos Athanasiou and
Michael J. Black},
title = {{VIBE}: Video Inference for Human Body Pose and Shape Estimation},
booktitle = {CVPR},
year = {2020}
}
- SMPL v1.0 is used in our experiments.
- J_regressor_extra.npy
- J_regressor_h36m.npy
- smpl_mean_params.npz
- The pretrained frame feature extractor spin.pth
Download the above resources and arrange them in the following file structure:
mmhuman3d
├── mmhuman3d
├── docs
├── tests
├── tools
├── configs
└── data
├── body_models
│ ├── J_regressor_extra.npy
│ ├── J_regressor_h36m.npy
│ ├── smpl_mean_params.npz
│ └── smpl
│ ├── SMPL_FEMALE.pkl
│ ├── SMPL_MALE.pkl
│ └── SMPL_NEUTRAL.pkl
├── pretrained
│ └── spin.pth
├── preprocessed_datasets
│ ├── vibe_mpi_inf_3dhp_train.npz
│ └── vibe_insta_variety.npz
└── datasets
└── mpi_inf_3dhp
We evaluate VIBE on 3DPW. Values are MPJPE/PA-MPJPE.
Config | 3DPW | Download |
---|---|---|
resnet50_vibe_pw3d.py | 94.89 / 57.08 | model | log |