Enhancing Diffusion Model Stability for Image Restoration via Gradient Management
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909767847378944 |
|---|---|
| author | Wu, Hongjie Zhang, Mingqin He, Linchao Zhou, Ji-Zhe Lv, Jiancheng |
| author_facet | Wu, Hongjie Zhang, Mingqin He, Linchao Zhou, Ji-Zhe Lv, Jiancheng |
| contents | Diffusion models have shown remarkable promise for image restoration by leveraging powerful priors. Prominent methods typically frame the restoration problem within a Bayesian inference framework, which iteratively combines a denoising step with a likelihood guidance step. However, the interactions between these two components in the generation process remain underexplored. In this paper, we analyze the underlying gradient dynamics of these components and identify significant instabilities. Specifically, we demonstrate conflicts between the prior and likelihood gradient directions, alongside temporal fluctuations in the likelihood gradient itself. We show that these instabilities disrupt the generative process and compromise restoration performance. To address these issues, we propose Stabilized Progressive Gradient Diffusion (SPGD), a novel gradient management technique. SPGD integrates two synergistic components: (1) a progressive likelihood warm-up strategy to mitigate gradient conflicts; and (2) adaptive directional momentum (ADM) smoothing to reduce fluctuations in the likelihood gradient. Extensive experiments across diverse restoration tasks demonstrate that SPGD significantly enhances generation stability, leading to state-of-the-art performance in quantitative metrics and visually superior results. Code is available at https://github.com/74587887/SPGD. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_06656 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Enhancing Diffusion Model Stability for Image Restoration via Gradient Management Wu, Hongjie Zhang, Mingqin He, Linchao Zhou, Ji-Zhe Lv, Jiancheng Computer Vision and Pattern Recognition Machine Learning Diffusion models have shown remarkable promise for image restoration by leveraging powerful priors. Prominent methods typically frame the restoration problem within a Bayesian inference framework, which iteratively combines a denoising step with a likelihood guidance step. However, the interactions between these two components in the generation process remain underexplored. In this paper, we analyze the underlying gradient dynamics of these components and identify significant instabilities. Specifically, we demonstrate conflicts between the prior and likelihood gradient directions, alongside temporal fluctuations in the likelihood gradient itself. We show that these instabilities disrupt the generative process and compromise restoration performance. To address these issues, we propose Stabilized Progressive Gradient Diffusion (SPGD), a novel gradient management technique. SPGD integrates two synergistic components: (1) a progressive likelihood warm-up strategy to mitigate gradient conflicts; and (2) adaptive directional momentum (ADM) smoothing to reduce fluctuations in the likelihood gradient. Extensive experiments across diverse restoration tasks demonstrate that SPGD significantly enhances generation stability, leading to state-of-the-art performance in quantitative metrics and visually superior results. Code is available at https://github.com/74587887/SPGD. |
| title | Enhancing Diffusion Model Stability for Image Restoration via Gradient Management |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2507.06656 |