Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -33,3 +33,36 @@ configs:
|
|
33 |
- split: train
|
34 |
path: data/train-*
|
35 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
- split: train
|
34 |
path: data/train-*
|
35 |
---
|
36 |
+
|
37 |
+
# SA-Text
|
38 |
+
|
39 |
+
**Text-Aware Image Restoration with Diffusion Models** (arXiv:2506.09993)
|
40 |
+
Large-scale training dataset for the **Text-Aware Image Restoration (TAIR)** task.
|
41 |
+
|
42 |
+
- 📄 Paper: https://arxiv.org/abs/2506.09993
|
43 |
+
- 🌐 Project Page: https://cvlab-kaist.github.io/TAIR/
|
44 |
+
- 💻 GitHub: https://github.com/cvlab-kaist/TAIR
|
45 |
+
- 🛠 Dataset Pipeline: https://github.com/paulcho98/text_restoration_dataset
|
46 |
+
|
47 |
+
## Dataset Description
|
48 |
+
|
49 |
+
**SA-Text** is constructed from SA-1B dataset using our official [dataset pipeline](https://github.com/paulcho98/text_restoration_dataset). It contains **100K** high-resolution scene images paired with polygon-level text annotations.
|
50 |
+
This dataset is tailored for TAIR task, which aims to restore both visual quality and text fidelity in degraded images.
|
51 |
+
|
52 |
+
## Notes
|
53 |
+
|
54 |
+
- Each image includes one or more **text instances** with transcriptions and polygon-level labels.
|
55 |
+
- Designed for training **TeReDiff**, a multi-task diffusion model introduced in our paper.
|
56 |
+
- For real-world evaluation, check [Real-Text](https://huggingface.co/datasets/Min-Jaewon/Real-Text).
|
57 |
+
|
58 |
+
## Citation
|
59 |
+
Please cite the following paper if you use this dataset:
|
60 |
+
```
|
61 |
+
{
|
62 |
+
@article{min2024textaware,
|
63 |
+
title={Text-Aware Image Restoration with Diffusion Models},
|
64 |
+
author={Min, Jaewon and Kim, Jin Hyeon and Cho, Paul Hyunbin and Lee, Jaeeun and Park, Jihye and Park, Minkyu and Kim, Sangpil and Park, Hyunhee and Kim, Seungryong},
|
65 |
+
journal={arXiv preprint arXiv:2506.09993},
|
66 |
+
year={2025}
|
67 |
+
}
|
68 |
+
```
|