D$^2$-VR: Degradation-Robust and Distilled Video Restoration with Synergistic Optimization Strategy

2026-02-09

AI Summary

This work addresses the challenges of high inference latency and temporal inconsistency in existing diffusion prior–based video restoration methods when handling complex real-world degradations. The authors propose a single-image diffusion-driven, low-step video restoration framework that leverages a degradation-robust optical flow alignment module to obtain reliable temporal guidance. By integrating adversarial distillation to compress the diffusion sampling steps and introducing a co-optimization strategy to jointly enhance perceptual quality and temporal consistency, the method achieves state-of-the-art restoration performance while accelerating diffusion sampling by 12× compared to current approaches.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Adversarial Learning & Robustness
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

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$\times$}

Problem

• video restoration
• inference latency
• temporal instability
• real-world degradations
• diffusion priors

Innovation

• Degradation-Robust Flow Alignment
• Adversarial Distillation
• Diffusion-based Video Restoration
• Synergistic Optimization
• Low-step Inference