SmoothNet (arXiv'2021)
@article{zeng2021smoothnet,
title={SmoothNet: A Plug-and-Play Network for Refining Human Poses in Videos},
author={Zeng, Ailing and Yang, Lei and Ju, Xuan and Li, Jiefeng and Wang, Jianyi and Xu, Qiang},
journal={arXiv preprint arXiv:2112.13715},
year={2021}
}
Human3.6M (TPAMI'2014)
@article{h36m_pami,
author = {Ionescu, Catalin and Papava, Dragos and Olaru, Vlad and Sminchisescu, Cristian},
title = {Human3.6M: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
publisher = {IEEE Computer Society},
volume = {36},
number = {7},
pages = {1325-1339},
month = {jul},
year = {2014}
}
The following SmoothNet model checkpoints are available for pose smoothing. The table shows the the performance of SimpleBaseline3D on Human3.6M dataset without/with the SmoothNet plugin, and compares the SmoothNet models with 4 different window sizes (8, 16, 32 and 64). The metrics are MPJPE(mm), P-MEJPE(mm) and Acceleration Error (mm/frame^2).
| Arch | Window Size | MPJPEw/o | MPJPEw | P-MPJPEw/o | P-MPJPEw | AC. Errw/o | AC. Errw | ckpt |
|---|---|---|---|---|---|---|---|---|
| smoothnet_ws8 | 8 | 54.48 | 53.15 | 42.20 | 41.32 | 19.18 | 1.87 | ckpt |
| smoothnet_ws16 | 16 | 54.48 | 52.74 | 42.20 | 41.20 | 19.18 | 1.22 | ckpt |
| smoothnet_ws32 | 32 | 54.48 | 52.47 | 42.20 | 40.84 | 19.18 | 0.99 | ckpt |
| smoothnet_ws64 | 64 | 54.48 | 53.37 | 42.20 | 40.77 | 19.18 | 0.92 | ckpt |