README.md

---
license: cc-by-nc-sa-4.0
size_categories:
- 1K<n<10K
task_categories:
- keypoint-detection
- pose-estimation
---
[Paper](https://arxiv.org/pdf/2407.13930)

# RT-Pose: A 4D Radar Tensor-based 3D Human Pose Estimation and Localization Benchmark (ECCV 2024)

RT-Pose introduces a human pose estimation (HPE) dataset and benchmark by integrating a unique combination of calibrated radar ADC data, 4D radar tensors, stereo RGB images, and LiDAR point clouds. 
This integration marks a significant advancement in studying human pose analysis through multi-modality datasets.

![images](./asset/data_viz.gif)
![images](./asset/annotation.gif)


## Dataset Details

### Dataset Description

<!-- Provide a longer summary of what this dataset is. -->
#### Sensors
The data collection hardware system comprises two RGB [cameras](https://www.flir.com/products/blackfly-s-usb3/?model=BFS-U3-16S2C-CS), a non-repetitive
horizontal scanning [LiDAR](https://www.livoxtech.com/3296f540ecf5458a8829e01cf429798e/assets/horizon/Livox%20Horizon%20user%20manual%20v1.0.pdf), and a cascade imaging [radar module](https://www.ti.com/tool/MMWCAS-RF-EVM).
![images](./asset/device.png)


#### Data Statics
We collect the dataset in 40 scenes with indoor and outdoor environments.
![images](./asset/examples.png)


The dataset comprises 72,000 frames distributed across 240 sequences. 
The structured organization ensures a realistic distribution of human motions, which is crucial for robust analysis and model training.

![images](./asset/data_distribution.png)

Please check the paper for more details.

- **Curated by:** Yuan-Hao Ho (n28081527@gs.ncku.edu.tw), Jen-Hao(Andy) Cheng(andyhci@uw.edu) from [Information Processing Lab](https://ipl-uw.github.io/) at University of Washington
- **License:** [CC BY-NC-SA](https://creativecommons.org/licenses/by-nc-sa/4.0/deed.en)

### Dataset Sources

<!-- Provide the basic links for the dataset. -->

- **Repository including data processing and baseline method codes:** [RT-POSE](https://github.com/ipl-uw/RT-POSE)
- **Paper:** [Paper](https://arxiv.org/pdf/2407.13930)

## Uses

<!-- Address questions around how the dataset is intended to be used. -->
1. Download the dataset from Hugging Face (Total data size: ~1.2 TB)
2. Follow the [data processing tool](https://github.com/ipl-uw/RT-POSE/data_processing) to process radar ADC samples into radar tensors. (Total data size of the downloaded data and saved radar tensors: ~41 TB)
3. Check the data loading and baseline method's training and testing codes in the same repo [RT-POSE](https://github.com/ipl-uw/RT-POSE)

## Citation

**BibTeX:**

@article{rtpose2024,
      title={RT-Pose: A 4D Radar Tensor-based 3D Human Pose Estimation and Localization Benchmark}, 
      author={Yuan-Hao Ho and Jen-Hao Cheng and Sheng Yao Kuan and Zhongyu Jiang and Wenhao Chai and Hsiang-Wei Huang and Chih-Lung Lin and Jenq-Neng Hwang},
      journal={arXiv preprint arXiv:2407.13930},
      year={2024}
}