Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,64 @@
|
|
1 |
---
|
2 |
-
license: cc-by-nc-
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
license: cc-by-nc-sa-4.0
|
3 |
+
size_categories:
|
4 |
+
- 1K<n<10K
|
5 |
---
|
6 |
+
|
7 |
+
# RT-Pose: A 4D Radar Tensor-based 3D Human Pose Estimation and Localization Benchmark (ECCV 2024)
|
8 |
+
|
9 |
+
RT-Pose introduces a human pose estimation (HPE) dataset and benchmark by integrating a unique combination of calibrated 4D radar tensors, RGB images, and LiDAR point clouds.
|
10 |
+
This integration marks a significant advancement in studying human pose analysis through multi-modality datasets.
|
11 |
+
|
12 |
+
|
13 |
+
|
14 |
+
## Dataset Details
|
15 |
+
|
16 |
+
### Dataset Description
|
17 |
+
|
18 |
+
<!-- Provide a longer summary of what this dataset is. -->
|
19 |
+
#### Sensors
|
20 |
+
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
|
21 |
+
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).
|
22 |
+
![images](./asset/device.pdf)
|
23 |
+
|
24 |
+
|
25 |
+
#### Data Statics
|
26 |
+
We collect the dataset in 40 scenes with indoor and outdoor environments.
|
27 |
+
![images](./asset/examples.pdf)
|
28 |
+
|
29 |
+
|
30 |
+
The dataset comprises 72,000 frames distributed across 240 sequences.
|
31 |
+
The structured organization ensures a realistic distribution of human motions, which is crucial for robust analysis and model training.
|
32 |
+
|
33 |
+
![images](./asset/data_distribution.pdf)
|
34 |
+
|
35 |
+
Please check the paper for more details.
|
36 |
+
|
37 |
+
- **Curated by:** Jen-Hao(Andy) Cheng(andyhci@uw.edu), Yuan-Hao Ho (n28081527@gs.ncku.edu.tw) from [Information Processing Lab](https://ipl-uw.github.io/) at University of Washington
|
38 |
+
- **License:** [CC BY-NC-SA](https://creativecommons.org/licenses/by-nc-sa/4.0/deed.en)
|
39 |
+
|
40 |
+
### Dataset Sources
|
41 |
+
|
42 |
+
<!-- Provide the basic links for the dataset. -->
|
43 |
+
|
44 |
+
- **Repository including data processing and baseline method codes:** [RT-POSE](https://github.com/ipl-uw/RT-POSE)
|
45 |
+
- **Paper:** To be viewed on arxiv.
|
46 |
+
|
47 |
+
## Uses
|
48 |
+
|
49 |
+
<!-- Address questions around how the dataset is intended to be used. -->
|
50 |
+
1. Download the dataset from Hugging Face (Total data size: ~1.2 TB)
|
51 |
+
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)
|
52 |
+
3. Check the data loading and baseline training, testing codes in the same repo [RT-POSE](https://github.com/ipl-uw/RT-POSE)
|
53 |
+
|
54 |
+
## Citation
|
55 |
+
|
56 |
+
<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->
|
57 |
+
|
58 |
+
**BibTeX:**
|
59 |
+
|
60 |
+
To appear on arxiv
|
61 |
+
|
62 |
+
## Dataset Card Contact
|
63 |
+
|
64 |
+
Jen-Hao (Andy) Cheng, andyhci@uw.edu
|