Deep High-Resolution Representation Learning for Human Pose Estimation (accepted to CVPR2019)

News

• If you are interested in internship or research positions related to computer vision in ByteDance AI Lab, feel free to contact me(leoxiaobin-at-gmail.com).
• Our new work High-Resolution Representations for Labeling Pixels and Regions is available at HRNet. Our HRNet has been applied to a wide range of vision tasks, such as image classification, objection detection, semantic segmentation and facial landmark.

Introduction

This is an official pytorch implementation of Deep High-Resolution Representation Learning for Human Pose Estimation. In this work, we are interested in the human pose estimation problem with a focus on learning reliable high-resolution representations. Most existing methods recover high-resolution representations from low-resolution representations produced by a high-to-low resolution network. Instead, our proposed network maintains high-resolution representations through the whole process. We start from a high-resolution subnetwork as the first stage, gradually add high-to-low resolution subnetworks one by one to form more stages, and connect the mutli-resolution subnetworks in parallel. We conduct repeated multi-scale fusions such that each of the high-to-low resolution representations receives information from other parallel representations over and over, leading to rich high-resolution representations. As a result, the predicted keypoint heatmap is potentially more accurate and spatially more precise. We empirically demonstrate the effectiveness of our network through the superior pose estimation results over two benchmark datasets: the COCO keypoint detection dataset and the MPII Human Pose dataset.

Main Results

Results on MPII val

Note:

• Flip test is used.
• Input size is 256x256
• pose_resnet_[50,101,152] is our previous work of Simple Baselines for Human Pose Estimation and Tracking

Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset

Note:

• Flip test is used.
• Person detector has person AP of 56.4 on COCO val2017 dataset.
• pose_resnet_[50,101,152] is our previous work of Simple Baselines for Human Pose Estimation and Tracking.
• GFLOPs is for convolution and linear layers only.

Results on COCO test-dev2017 with detector having human AP of 60.9 on COCO test-dev2017 dataset

Note:

• Flip test is used.
• Person detector has person AP of 60.9 on COCO test-dev2017 dataset.
• pose_resnet_152 is our previous work of Simple Baselines for Human Pose Estimation and Tracking.
• GFLOPs is for convolution and linear layers only.
• pose_hrnet_w48* means using additional data from AI challenger for training.

Environment

The code is developed using python 3.6 on Ubuntu 16.04. NVIDIA GPUs are needed. The code is developed and tested using 4 NVIDIA P100 GPU cards. Other platforms or GPU cards are not fully tested.

Quick start

Installation

1. Install pytorch >= v1.0.0 following official instruction. Note that if you use pytorch's version < v1.0.0, you should following the instruction at https://github.com/Microsoft/human-pose-estimation.pytorch to disable cudnn's implementations of BatchNorm layer. We encourage you to use higher pytorch's version(>=v1.0.0)

2. Clone this repo, and we'll call the directory that you cloned as ${POSE_ROOT}.

3. Install dependencies:

   pip install -r requirements.txt
   

4. Make libs:

   cd ${POSE_ROOT}/lib
   make
   

5. Install COCOAPI:

   # COCOAPI=/path/to/clone/cocoapi
   git clone https://github.com/cocodataset/cocoapi.git $COCOAPI
   cd $COCOAPI/PythonAPI
   # Install into global site-packages
   make install
   # Alternatively, if you do not have permissions or prefer
   # not to install the COCO API into global site-packages
   python3 setup.py install --user
   

   Note that instructions like # COCOAPI=/path/to/install/cocoapi indicate that you should pick a path where you'd like to have the software cloned and then set an environment variable (COCOAPI in this case) accordingly.

6. Init output(training model output directory) and log(tensorboard log directory) directory:

   mkdir output 
   mkdir log
   

   Your directory tree should look like this:

   ${POSE_ROOT}
   ├── data
   ├── experiments
   ├── lib
   ├── log
   ├── models
   ├── output
   ├── tools 
   ├── README.md
   └── requirements.txt
   

7. Download pretrained models from our model zoo(GoogleDrive or OneDrive)

   ${POSE_ROOT}
    `-- models
        `-- pytorch
            |-- imagenet
            |   |-- hrnet_w32-36af842e.pth
            |   |-- hrnet_w48-8ef0771d.pth
            |   |-- resnet50-19c8e357.pth
            |   |-- resnet101-5d3b4d8f.pth
            |   `-- resnet152-b121ed2d.pth
            |-- pose_coco
            |   |-- pose_hrnet_w32_256x192.pth
            |   |-- pose_hrnet_w32_384x288.pth
            |   |-- pose_hrnet_w48_256x192.pth
            |   |-- pose_hrnet_w48_384x288.pth
            |   |-- pose_resnet_101_256x192.pth
            |   |-- pose_resnet_101_384x288.pth
            |   |-- pose_resnet_152_256x192.pth
            |   |-- pose_resnet_152_384x288.pth
            |   |-- pose_resnet_50_256x192.pth
            |   `-- pose_resnet_50_384x288.pth
            `-- pose_mpii
                |-- pose_hrnet_w32_256x256.pth
                |-- pose_hrnet_w48_256x256.pth
                |-- pose_resnet_101_256x256.pth
                |-- pose_resnet_152_256x256.pth
                `-- pose_resnet_50_256x256.pth
   
   

Data preparation

For MPII data, please download from MPII Human Pose Dataset. The original annotation files are in matlab format. We have converted them into json format, you also need to download them from OneDrive or GoogleDrive. Extract them under {POSE_ROOT}/data, and make them look like this:

${POSE_ROOT}
|-- data
`-- |-- mpii
    `-- |-- annot
        |   |-- gt_valid.mat
        |   |-- test.json
        |   |-- train.json
        |   |-- trainval.json
        |   `-- valid.json
        `-- images
            |-- 000001163.jpg
            |-- 000003072.jpg

For COCO data, please download from COCO download, 2017 Train/Val is needed for COCO keypoints training and validation. We also provide person detection result of COCO val2017 and test-dev2017 to reproduce our multi-person pose estimation results. Please download from OneDrive or GoogleDrive. Download and extract them under {POSE_ROOT}/data, and make them look like this:

${POSE_ROOT}
|-- data
`-- |-- coco
    `-- |-- annotations
        |   |-- person_keypoints_train2017.json
        |   `-- person_keypoints_val2017.json
        |-- person_detection_results
        |   |-- COCO_val2017_detections_AP_H_56_person.json
        |   |-- COCO_test-dev2017_detections_AP_H_609_person.json
        `-- images
            |-- train2017
            |   |-- 000000000009.jpg
            |   |-- 000000000025.jpg
            |   |-- 000000000030.jpg
            |   |-- ... 
            `-- val2017
                |-- 000000000139.jpg
                |-- 000000000285.jpg
                |-- 000000000632.jpg
                |-- ... 

Training and Testing

Testing on MPII dataset using model zoo's models(GoogleDrive or OneDrive)

python tools/test.py \
    --cfg experiments/mpii/hrnet/w32_256x256_adam_lr1e-3.yaml \
    TEST.MODEL_FILE models/pytorch/pose_mpii/pose_hrnet_w32_256x256.pth

Training on MPII dataset

python tools/train.py \
    --cfg experiments/mpii/hrnet/w32_256x256_adam_lr1e-3.yaml

Testing on COCO val2017 dataset using model zoo's models(GoogleDrive or OneDrive)

python tools/test.py \
    --cfg experiments/coco/hrnet/w32_256x192_adam_lr1e-3.yaml \
    TEST.MODEL_FILE models/pytorch/pose_coco/pose_hrnet_w32_256x192.pth \
    TEST.USE_GT_BBOX False

Training on COCO train2017 dataset

python tools/train.py \
    --cfg experiments/coco/hrnet/w32_256x192_adam_lr1e-3.yaml \

Other applications

Many other dense prediction tasks, such as segmentation, face alignment and object detection, etc. have been benefited by HRNet. More information can be found at Deep High-Resolution Representation Learning.

Citation

If you use our code or models in your research, please cite with:

@inproceedings{sun2019deep,
  title={Deep High-Resolution Representation Learning for Human Pose Estimation},
  author={Sun, Ke and Xiao, Bin and Liu, Dong and Wang, Jingdong},
  booktitle={CVPR},
  year={2019}
}

@inproceedings{xiao2018simple,
    author={Xiao, Bin and Wu, Haiping and Wei, Yichen},
    title={Simple Baselines for Human Pose Estimation and Tracking},
    booktitle = {European Conference on Computer Vision (ECCV)},
    year = {2018}
}