Face Models

Distribution-aware coordinate representation for human pose estimation

Introduction

@inproceedings{zhang2020distribution,
  title={Distribution-aware coordinate representation for human pose estimation},
  author={Zhang, Feng and Zhu, Xiatian and Dai, Hanbin and Ye, Mao and Zhu, Ce},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={7093--7102},
  year={2020}
}

Results and models

2d Face Keypoint Estimation

Results on AFLW dataset

The model is trained on AFLW train and evaluated on AFLW full and frontal.

Arch	Input Size	NMEfull	NMEfrontal	ckpt	log
dark_pose_hrnetv2_w18	256x256	1.41	1.27	ckpt	log

Results on WFLW dataset

The model is trained on WFLW train.

Arch	Input Size	NMEtest	NMEpose	NMEillumination	NMEocclusion	NMEblur	NMEmakeup	NMEexpression	ckpt	log
dark_pose_hrnetv2_w18	256x256	3.98	6.99	3.96	4.78	4.57	3.87	4.30	ckpt	log

Deeppose + Wingloss

Introduction

@inproceedings{toshev2014deeppose,
  title={Deeppose: Human pose estimation via deep neural networks},
  author={Toshev, Alexander and Szegedy, Christian},
  booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition},
  pages={1653--1660},
  year={2014}
}

@inproceedings{feng2018wing,
  title={Wing Loss for Robust Facial Landmark Localisation with Convolutional Neural Networks},
  author={Feng, Zhen-Hua and Kittler, Josef and Awais, Muhammad and Huber, Patrik and Wu, Xiao-Jun},
  booktitle={Computer Vision and Pattern Recognition (CVPR), 2018 IEEE Conference on},
  year={2018},
  pages ={2235-2245},
  organization={IEEE}
}

Results and models

2d Face Keypoint Estimation

Results on WFLW dataset

The model is trained on WFLW train.

Arch	Input Size	NMEtest	NMEpose	NMEillumination	NMEocclusion	NMEblur	NMEmakeup	NMEexpression	ckpt	log
deeppose_res50	256x256	4.85	8.50	4.81	5.69	5.45	4.82	5.20	ckpt	log
deeppose_res50_wingloss	256x256	4.64	8.25	4.59	5.56	5.26	4.59	5.07	ckpt	log

Deep high-resolution representation learning for visual recognition

Introduction

@article{WangSCJDZLMTWLX19,
  title={Deep High-Resolution Representation Learning for Visual Recognition},
  author={Jingdong Wang and Ke Sun and Tianheng Cheng and
          Borui Jiang and Chaorui Deng and Yang Zhao and Dong Liu and Yadong Mu and
          Mingkui Tan and Xinggang Wang and Wenyu Liu and Bin Xiao},
  journal   = {TPAMI}
  year={2019}
}

Results and models

2d Face Keypoint Estimation

Results on AFLW dataset

The model is trained on AFLW train and evaluated on AFLW full and frontal.

Arch	Input Size	NMEfull	NMEfrontal	ckpt	log
pose_hrnetv2_w18	256x256	1.41	1.27	ckpt	log

Results on WFLW dataset

The model is trained on WFLW train.

Arch	Input Size	NMEtest	NMEpose	NMEillumination	NMEocclusion	NMEblur	NMEmakeup	NMEexpression	ckpt	log
pose_hrnetv2_w18	256x256	4.06	6.98	3.99	4.83	4.59	3.92	4.33	ckpt	log

Simple baselines for human pose estimation and tracking

Introduction

@inproceedings{xiao2018simple,
  title={Simple baselines for human pose estimation and tracking},
  author={Xiao, Bin and Wu, Haiping and Wei, Yichen},
  booktitle={Proceedings of the European conference on computer vision (ECCV)},
  pages={466--481},
  year={2018}
}