D$^2$-VR: Degradation-Robust and Distilled Video Restoration with Synergistic Optimization Strategy
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910016403931136 |
|---|---|
| author | Liang, Jianfeng Shen, Shaocheng Xu, Botao Hu, Qiang Zhang, Xiaoyun |
| author_facet | Liang, Jianfeng Shen, Shaocheng Xu, Botao Hu, Qiang Zhang, Xiaoyun |
| contents | The integration of diffusion priors with temporal alignment has emerged as a transformative paradigm for video restoration, delivering fantastic perceptual quality, yet the practical deployment of such frameworks is severely constrained by prohibitive inference latency and temporal instability when confronted with complex real-world degradations. To address these limitations, we propose \textbf{D$^2$-VR}, a single-image diffusion-based video-restoration framework with low-step inference. To obtain precise temporal guidance under severe degradation, we first design a Degradation-Robust Flow Alignment (DRFA) module that leverages confidence-aware attention to filter unreliable motion cues. We then incorporate an adversarial distillation paradigm to compress the diffusion sampling trajectory into a rapid few-step regime. Finally, a synergistic optimization strategy is devised to harmonize perceptual quality with rigorous temporal consistency. Extensive experiments demonstrate that D$^2$-VR achieves state-of-the-art performance while accelerating the sampling process by \textbf{12$\times$} |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08395 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | D$^2$-VR: Degradation-Robust and Distilled Video Restoration with Synergistic Optimization Strategy Liang, Jianfeng Shen, Shaocheng Xu, Botao Hu, Qiang Zhang, Xiaoyun Computer Vision and Pattern Recognition The integration of diffusion priors with temporal alignment has emerged as a transformative paradigm for video restoration, delivering fantastic perceptual quality, yet the practical deployment of such frameworks is severely constrained by prohibitive inference latency and temporal instability when confronted with complex real-world degradations. To address these limitations, we propose \textbf{D$^2$-VR}, a single-image diffusion-based video-restoration framework with low-step inference. To obtain precise temporal guidance under severe degradation, we first design a Degradation-Robust Flow Alignment (DRFA) module that leverages confidence-aware attention to filter unreliable motion cues. We then incorporate an adversarial distillation paradigm to compress the diffusion sampling trajectory into a rapid few-step regime. Finally, a synergistic optimization strategy is devised to harmonize perceptual quality with rigorous temporal consistency. Extensive experiments demonstrate that D$^2$-VR achieves state-of-the-art performance while accelerating the sampling process by \textbf{12$\times$} |
| title | D$^2$-VR: Degradation-Robust and Distilled Video Restoration with Synergistic Optimization Strategy |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.08395 |