Update README.md
Browse files
README.md
CHANGED
|
@@ -28,26 +28,6 @@ This Hugging Face repository is intended to provide:
|
|
| 28 |
|
| 29 |
---
|
| 30 |
|
| 31 |
-
## Supported checkpoints
|
| 32 |
-
|
| 33 |
-
You can organize the released weights in this repository with names similar to the following:
|
| 34 |
-
|
| 35 |
-
| Checkpoint | Task | Dataset |
|
| 36 |
-
|---|---|---|
|
| 37 |
-
| `resfusion-istd.ckpt` | Shadow removal | ISTD |
|
| 38 |
-
| `resfusion-lol.ckpt` | Low-light enhancement | LOL |
|
| 39 |
-
| `resfusion-raindrop.ckpt` | Raindrop removal | Raindrop |
|
| 40 |
-
|
| 41 |
-
---
|
| 42 |
-
|
| 43 |
-
## Model details
|
| 44 |
-
|
| 45 |
-
- **Model type:** DDPM-based image restoration
|
| 46 |
-
- **Framework:** PyTorch Lightning
|
| 47 |
-
- **Primary use:** Paired image restoration
|
| 48 |
-
- **Input:** Degraded image
|
| 49 |
-
- **Output:** Restored image
|
| 50 |
-
|
| 51 |
### Intended uses
|
| 52 |
|
| 53 |
This model is intended for research and experimental applications in:
|
|
@@ -97,10 +77,15 @@ Typical workflow:
|
|
| 97 |
If you use Resfusion in your work, please cite:
|
| 98 |
|
| 99 |
```bibtex
|
| 100 |
-
@inproceedings{
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
}
|
| 106 |
```
|
|
|
|
| 28 |
|
| 29 |
---
|
| 30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
### Intended uses
|
| 32 |
|
| 33 |
This model is intended for research and experimental applications in:
|
|
|
|
| 77 |
If you use Resfusion in your work, please cite:
|
| 78 |
|
| 79 |
```bibtex
|
| 80 |
+
@inproceedings{NEURIPS2024_ebc62a3a,
|
| 81 |
+
author = {Shi, Zhenning and zheng, haoshuai and Xu, Chen and Dong, Changsheng and Pan, Bin and xueshuo, Xie and He, Along and Li, Tao and Fu, Huazhu},
|
| 82 |
+
booktitle = {Advances in Neural Information Processing Systems},
|
| 83 |
+
editor = {A. Globerson and L. Mackey and D. Belgrave and A. Fan and U. Paquet and J. Tomczak and C. Zhang},
|
| 84 |
+
pages = {130664--130693},
|
| 85 |
+
publisher = {Curran Associates, Inc.},
|
| 86 |
+
title = {Resfusion: Denoising Diffusion Probabilistic Models for Image Restoration Based on Prior Residual Noise},
|
| 87 |
+
url = {https://proceedings.neurips.cc/paper_files/paper/2024/file/ebc62a3af9342eb4ebc728e5c5bc4cca-Paper-Conference.pdf},
|
| 88 |
+
volume = {37},
|
| 89 |
+
year = {2024}
|
| 90 |
}
|
| 91 |
```
|