---
title: human_wholebody_keypoint_dataset
canonical_url: "https://www.modelscope.cn/datasets/modelscope/human_wholebody_keypoint_dataset"
md_url: "https://www.modelscope.cn/datasets/modelscope/human_wholebody_keypoint_dataset.md"
repository: modelscope/human_wholebody_keypoint_dataset
chinese_name: "全身关键点检测数据集"
last_updated: 2022-10-18
license: "Apache License 2.0"
storage_size: "4.7 MB"
domain:
  - image
tasks:
  - human-wholebody-keypoints
color_space:
  - "彩色图片"
size_scale:
  - 10k-1m
downloads: 957
stars: 5
---

# human_wholebody_keypoint_dataset

> human_wholebody_keypoint_dataset - modelscope 在 ModelScope 开源的数据集。本数据集来源于coco，包含训练集、测试集和训练子集三部分，主要用于133点全身关键点模型训练和测试。

modelscope/human_wholebody_keypoint_dataset 是 ModelScope 魔搭社区上的human-wholebody-keypoints数据集，涉及 image 领域，存储大小 4.7 MB，采用 Apache License 2.0 许可。

- **Repository**: modelscope/human_wholebody_keypoint_dataset
- **License**: Apache License 2.0
- **Tasks**: human-wholebody-keypoints
- **Domain**: image
- **Storage size**: 4.7 MB
- **Downloads**: 957
- **Stars**: 5
- **Last updated**: 2022-10-18

Source: https://www.modelscope.cn/datasets/modelscope/human_wholebody_keypoint_dataset

---

## 数据集描述
本数据集为全身关键点检测数据集，用于全身关键点检测模型的训练和测试，本数据集来源于公开数据集 [COCO-WholeBody](https://github.com/jin-s13/COCO-WholeBody) 。

### 数据集简介
本数据集是[COCO-WholeBody](https://github.com/jin-s13/COCO-WholeBody) 。COCO-WholeBody是COCO 2017 dataset的扩展，保持同样的train和val数据集划分.

### 数据集支持的任务
本数据集支持全身关键点检测模型的训练和测试任务。

## 数据集的格式和结构
### 数据格式
[标注格式参考这里。](https://github.com/jin-s13/COCO-WholeBody/blob/master/data_format.md)

<div align="center">
  <img src="./assets/keypoints.png" width="2910" />
</div>

### 数据集加载方式
本仓库目前提供了训练集、验证集和一个训练子集的下载。
```python
from modelscope.msdatasets import MsDataset
from modelscope.utils.constant import DownloadMode

# 训练集
dataset_train = MsDataset.load('human_wholebody_keypoint_dataset', namespace='damo', 
                               split='train',  download_mode=DownloadMode.FORCE_REDOWNLOAD)

# 训练子集，方便快速验证流程是否能够跑通
dataset_subtrain = MsDataset.load('human_wholebody_keypoint_dataset', namespace='damo', 
                                  split='subtrain',  download_mode=DownloadMode.FORCE_REDOWNLOAD)

# 验证集
dataset_val = MsDataset.load('human_wholebody_keypoint_dataset', namespace='damo', 
                             split='validation',  download_mode=DownloadMode.FORCE_REDOWNLOAD)

print(dataset_train.config_kwargs)
print(dataset_subtrain.config_kwargs)
print(dataset_val.config_kwargs)
```

### 数据分片
本数据集用于全身关键点检测模型训练和测试，提供`train`、`subtrain`和`validation`三个split。

## 数据集信息
本数据集主要来源于以下工作：

```BibTeX
@inproceedings{jin2020whole,
  title={Whole-Body Human Pose Estimation in the Wild},
  author={Jin, Sheng and Xu, Lumin and Xu, Jin and Wang, Can and Liu, Wentao and Qian, Chen and Ouyang, Wanli and Luo, Ping},
  booktitle={Proceedings of the European Conference on Computer Vision (ECCV)},    
  year={2020}
}

@article{xu2022zoomnas,
  title={ZoomNAS: Searching for Whole-body Human Pose Estimation in the Wild},
  author={Xu, Lumin and Jin, Sheng and Liu, Wentao and Qian, Chen and Ouyang, Wanli and Luo, Ping and Wang, Xiaogang},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year={2022},
  publisher={IEEE}
}
```

### Clone with HTTP
* http://www.modelscope.cn/datasets/modelscope/human_wholebody_keypoint_dataset.git