---
title: hand_detection_dataset
canonical_url: "https://www.modelscope.cn/datasets/modelscope/hand_detection_dataset"
md_url: "https://www.modelscope.cn/datasets/modelscope/hand_detection_dataset.md"
repository: modelscope/hand_detection_dataset
chinese_name: "手部检测数据集"
last_updated: 2022-12-15
license: "Apache License 2.0"
storage_size: "3.1 KB"
domain:
  - image
tasks:
  - hand-detection
size_scale:
  - 0-33000
downloads: 2889
stars: 6
---

# hand_detection_dataset

> hand_detection_dataset - modelscope 在 ModelScope 开源的数据集。本数据集来源于coco-hand-big和TV-hand，包含训练集和验证集两部分，已经转换成COCO格式，主要用于手部检测模型的训练和验证。

modelscope/hand_detection_dataset 是 ModelScope 魔搭社区上的hand-detection数据集，涉及 image 领域，存储大小 3.1 KB，采用 Apache License 2.0 许可。

- **Repository**: modelscope/hand_detection_dataset
- **License**: Apache License 2.0
- **Tasks**: hand-detection
- **Domain**: image
- **Storage size**: 3.1 KB
- **Downloads**: 2889
- **Stars**: 6
- **Last updated**: 2022-12-15

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

---

## 数据集描述
本数据集主要用于手部检测模型训练、测试和finetune演示，包含26000张训练图片和6500张验证图片，及其对应标注文件共计约60000个手部box，本数据集来自公开数据集TV-hand和coco-hand-big（https://www3.cs.stonybrook.edu/~cvl/projects/hand_det_attention/）。

### 数据集简介
由33000张图像组成，用于训练手部检测模型，其中26000张用于训练，6500张用于测试，，100张子集用于测试finetune代码。

### 数据集支持的任务
本数据集支持手部检测模型训练和finetune任务。

## 数据集的格式和结构

### 数据格式
数据集中包含训练和验证两部分，标注文件都已转换成COCO格式。

### 数据集加载方式
下面是加载数据集并使用手部检测模型进行测试的示例代码：

```python
from modelscope.msdatasets import MsDataset
from modelscope.utils.constant import DownloadMode

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

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

# 子集
dataset_subtrain = MsDataset.load('hand_detection_dataset', namespace='modelscope', 
                                  split='subtrain',  download_mode=DownloadMode.FORCE_REDOWNLOAD)

print(f'dataset_train.config_kwargs: {dataset_train.config_kwargs}')
print(f'dataset_val.config_kwargs: {dataset_val.config_kwargs}')
print(f'dataset_subtrain.config_kwargs: {dataset_subtrain.config_kwargs}')
```



## 数据集信息
本数据集主要来源于以下工作：

```
@article{Hand-CNN,
  title={Contextual Attention for Hand Detection in the Wild},
  author={Supreeth Narasimhaswamy and Zhengwei Wei and Yang Wang and Justin Zhang and Minh Hoai},
  booktitle={International Conference on Computer Vision (ICCV)},
  year={2019},
  url={https://arxiv.org/pdf/1904.04882.pdf} 
}
```

### Clone with HTTP
```bash
git clone https://www.modelscope.cn/datasets/modelscope/hand_detection_dataset.git
```
