Awesome

MXNet Multi Person Pose Estimation

This is a MXNet version of Realtime_Multi-Person_Pose_Estimation, original code is here https://github.com/ZheC/Realtime_Multi-Person_Pose_Estimation and here https://github.com/tensorboy/pytorch_Realtime_Multi-Person_Pose_Estimation

Introduction

Code repo for reproducing 2017 CVPR Oral paper using MXNet.

Require

1. MXNet
2. pip install mxnet-cu92 tensorflow mxboard
3. pip install pycocotools

Evalute

• python evaluate.py to evaluate the model on images separated by the original author
• It should have mAP 0.598 for the original rtpose, original repo have mAP 0.577 because we do left and right flip for heatmap and PAF for the evaluation.

Pretrained Models & Performance on the dataset split by the original rtpose.

Training

• cd data; bash getData.sh to obtain the COCO images in data/dataset/COCO/images/, keypoints annotations in data/dataset/COCO/annotations/
• Download the mask of the unlabeled person at Dropbox in data/dataset/COCO/mask
• Download the official training format at Dropbox in data/dataset/COCO.json

then run:

python train_model.py --gpu_ids 0 --lr=0.001 --wd=0.00001 --momentum=0.9 --log_key="lr_1_wd_0.0001_momentum_0.9"

Related repository

• CVPR'16, Convolutional Pose Machines.
• CVPR'17, Realtime Multi-Person Pose Estimation.

Network Architecture

• testing architecture

• training architecture

Contributions

All contributions are welcomed. If you encounter any issue (including examples of images where it fails) feel free to open an issue.

Citation

Please cite the paper in your publications if it helps your research:

@InProceedings{cao2017realtime,
  title = {Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields},
  author = {Zhe Cao and Tomas Simon and Shih-En Wei and Yaser Sheikh},
  booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2017}
  }
  
@INPROCEEDINGS{8486591, 
author={H. Wang and W. P. An and X. Wang and L. Fang and J. Yuan}, 
booktitle={2018 IEEE International Conference on Multimedia and Expo (ICME)}, 
title={Magnify-Net for Multi-Person 2D Pose Estimation}, 
year={2018}, 
volume={}, 
number={}, 
pages={1-6}, 
month={July},}