Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,62 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
<h2 align="center">Scale Space Diffusion [CVPR 2026]</h2>
|
| 6 |
+
|
| 7 |
+
<p align="center">
|
| 8 |
+
<a href="https://soumik-kanad.github.io/">Soumik Mukhopadhyay</a><sup>*</sup>,
|
| 9 |
+
<a href="https://prateksha.github.io/">Prateksha Udhayanan</a><sup>*</sup>,
|
| 10 |
+
<a href="https://abhinavsh.info/">Abhinav Shrivastava</a>
|
| 11 |
+
</p>
|
| 12 |
+
<p align="center">
|
| 13 |
+
<sup></sup>University of Maryland, College Park
|
| 14 |
+
<br>
|
| 15 |
+
<sup>*</sup>Equal contribution.
|
| 16 |
+
</p>
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
<p align="center">
|
| 20 |
+
<a href="https://prateksha.github.io/projects/scale-space-diffusion/"><img src="https://img.shields.io/badge/Website-Project_Page-2ea44f" /></a>
|
| 21 |
+
|
| 22 |
+
<a href="https://arxiv.org/abs/2603.08709"><img src="https://img.shields.io/badge/arXiv-2603.08709-b31b1b.svg" /></a>
|
| 23 |
+
|
| 24 |
+
<a href="https://github.com/prateksha/ScaleSpaceDiffusion"><img src="https://img.shields.io/badge/GitHub-Code-blue" /></a>
|
| 25 |
+
</p>
|
| 26 |
+
Scale Space Diffusion is a pixel-space diffusion model that integrates scale-space theory by combining Gaussian noise with linear degradations (downsampling). It introduces Flexi-UNet for resolution-aware denoising.
|
| 27 |
+
|
| 28 |
+
<table>
|
| 29 |
+
<tr>
|
| 30 |
+
<td align="center">
|
| 31 |
+
<img src="https://raw.githubusercontent.com/prateksha/ScaleSpaceDiffusion/main/assets/teaser.png"
|
| 32 |
+
alt="Scale Space Diffusion teaser" width="85%">
|
| 33 |
+
</td>
|
| 34 |
+
<td align="center">
|
| 35 |
+
<img src="https://raw.githubusercontent.com/prateksha/ScaleSpaceDiffusion/main/assets/ssd_gif.gif"
|
| 36 |
+
alt="Scale Space Diffusion sampling animation" width="100%">
|
| 37 |
+
</td>
|
| 38 |
+
</tr>
|
| 39 |
+
</table>
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
This model release accompanies the official implementation of **Scale Space Diffusion**.
|
| 43 |
+
|
| 44 |
+
## Model Details
|
| 45 |
+
|
| 46 |
+
The released checkpoints include SSD models trained on CelebA and ImageNet:
|
| 47 |
+
|
| 48 |
+
- CelebA-64: 2L and 4L Flexi-UNet checkpoints
|
| 49 |
+
- CelebA-128: 3L and 5L Flexi-UNet checkpoints
|
| 50 |
+
- CelebA-256: 3L and 6L Flexi-UNet checkpoints
|
| 51 |
+
- ImageNet-64: 2L Flexi-UNet checkpoint
|
| 52 |
+
|
| 53 |
+
## Citation
|
| 54 |
+
```
|
| 55 |
+
@article{mukhopadhyay2026scale,
|
| 56 |
+
title={Scale Space Diffusion},
|
| 57 |
+
author={Mukhopadhyay, Soumik and Udhayanan, Prateksha and Shrivastava, Abhinav},
|
| 58 |
+
journal={arXiv preprint arXiv:2603.08709},
|
| 59 |
+
year={2026}
|
| 60 |
+
}
|
| 61 |
+
```
|
| 62 |
+
|