Dynamic Attention-Guided Diffusion for Image Super-Resolution
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912319776227328 |
|---|---|
| author | Moser, Brian B. Frolov, Stanislav Raue, Federico Palacio, Sebastian Dengel, Andreas |
| author_facet | Moser, Brian B. Frolov, Stanislav Raue, Federico Palacio, Sebastian Dengel, Andreas |
| contents | Diffusion models in image Super-Resolution (SR) treat all image regions uniformly, which risks compromising the overall image quality by potentially introducing artifacts during denoising of less-complex regions. To address this, we propose ``You Only Diffuse Areas'' (YODA), a dynamic attention-guided diffusion process for image SR. YODA selectively focuses on spatial regions defined by attention maps derived from the low-resolution images and the current denoising time step. This time-dependent targeting enables a more efficient conversion to high-resolution outputs by focusing on areas that benefit the most from the iterative refinement process, i.e., detail-rich objects. We empirically validate YODA by extending leading diffusion-based methods SR3, DiffBIR, and SRDiff. Our experiments demonstrate new state-of-the-art performances in face and general SR tasks across PSNR, SSIM, and LPIPS metrics. As a side effect, we find that YODA reduces color shift issues and stabilizes training with small batches. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2308_07977 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Dynamic Attention-Guided Diffusion for Image Super-Resolution Moser, Brian B. Frolov, Stanislav Raue, Federico Palacio, Sebastian Dengel, Andreas Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning Diffusion models in image Super-Resolution (SR) treat all image regions uniformly, which risks compromising the overall image quality by potentially introducing artifacts during denoising of less-complex regions. To address this, we propose ``You Only Diffuse Areas'' (YODA), a dynamic attention-guided diffusion process for image SR. YODA selectively focuses on spatial regions defined by attention maps derived from the low-resolution images and the current denoising time step. This time-dependent targeting enables a more efficient conversion to high-resolution outputs by focusing on areas that benefit the most from the iterative refinement process, i.e., detail-rich objects. We empirically validate YODA by extending leading diffusion-based methods SR3, DiffBIR, and SRDiff. Our experiments demonstrate new state-of-the-art performances in face and general SR tasks across PSNR, SSIM, and LPIPS metrics. As a side effect, we find that YODA reduces color shift issues and stabilizes training with small batches. |
| title | Dynamic Attention-Guided Diffusion for Image Super-Resolution |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2308.07977 |