Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2605.24593 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914596467507200 |
|---|---|
| author | Hu, XiaoWan Yang, Jing Liu, HeNan Li, HuaQiu Xu, Mai |
| author_facet | Hu, XiaoWan Yang, Jing Liu, HeNan Li, HuaQiu Xu, Mai |
| contents | Zero-shot image restoration provides a flexible way to handle diverse degradations without task-specific training. However, existing methods typically rely on stacked layers or pre-trained features to enhance degradation expression, while overlooking physically consistent priors. The insufficient degradation prompts impose the heavy training burden and high sampling costs during zero-shot diffusion. Moreover, the fixed inference trajectory often collapses to suboptimal solutions under complex corruptions. We observe that heterogeneous degradations can be reparameterized into a minimal set of physically coherent parameters for compact representation. Based on this insight, we first propose a unified physical zero-shot image restoration (UP-ZeroIR) framework that explicitly models heterogeneous degradations into a homogeneous all-in-one distribution. The distribution can be optimized directly in the latent space, enabling principled solution exploration and effective prompt adaptation. Besides, we introduce a dynamic quality-refinement strategy that adaptively adjusts the diffusion trajectory for robust globally optimal convergence. Extensive experiments demonstrate that our method achieves state-of-the-art performance across both single and mixed degradations. Our code is available at https://github.com/yangjinglyy/UP-ZeroIR |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_24593 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Self-supervised Dynamic Heterogeneous Degradation Modeling for Unified Zero-Shot Image Restoration Hu, XiaoWan Yang, Jing Liu, HeNan Li, HuaQiu Xu, Mai Computer Vision and Pattern Recognition Zero-shot image restoration provides a flexible way to handle diverse degradations without task-specific training. However, existing methods typically rely on stacked layers or pre-trained features to enhance degradation expression, while overlooking physically consistent priors. The insufficient degradation prompts impose the heavy training burden and high sampling costs during zero-shot diffusion. Moreover, the fixed inference trajectory often collapses to suboptimal solutions under complex corruptions. We observe that heterogeneous degradations can be reparameterized into a minimal set of physically coherent parameters for compact representation. Based on this insight, we first propose a unified physical zero-shot image restoration (UP-ZeroIR) framework that explicitly models heterogeneous degradations into a homogeneous all-in-one distribution. The distribution can be optimized directly in the latent space, enabling principled solution exploration and effective prompt adaptation. Besides, we introduce a dynamic quality-refinement strategy that adaptively adjusts the diffusion trajectory for robust globally optimal convergence. Extensive experiments demonstrate that our method achieves state-of-the-art performance across both single and mixed degradations. Our code is available at https://github.com/yangjinglyy/UP-ZeroIR |
| title | Self-supervised Dynamic Heterogeneous Degradation Modeling for Unified Zero-Shot Image Restoration |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.24593 |