DUO-VSR: Dual-Stream Distillation for One-Step Video Super-Resolution

2026-03-23

AI Summary

This work addresses the challenges of training instability and insufficient supervision in existing single-step distillation methods for video super-resolution, as well as the high sampling cost of diffusion models. To this end, we propose a three-stage dual-stream distillation framework: it begins with progressive guidance distillation for initialization, followed by joint optimization of Distribution Matching Distillation (DMD) and adversarial supervision based on real/fake score features (RFS-GAN), and concludes with a preference-guided refinement mechanism. Our approach is the first to integrate dual-stream distillation with adversarial supervision, achieving superior visual quality compared to current single-step methods while significantly enhancing inference efficiency.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Diffusion-based video super-resolution (VSR) has recently achieved remarkable fidelity but still suffers from prohibitive sampling costs. While distribution matching distillation (DMD) can accelerate diffusion models toward one-step generation, directly applying it to VSR often results in training instability alongside degraded and insufficient supervision. To address these issues, we propose DUO-VSR, a three-stage framework built upon a Dual-Stream Distillation strategy that unifies distribution matching and adversarial supervision for one-step VSR. Firstly, a Progressive Guided Distillation Initialization is employed to stabilize subsequent training through trajectory-preserving distillation. Next, the Dual-Stream Distillation jointly optimizes the DMD and Real-Fake Score Feature GAN (RFS-GAN) streams, with the latter providing complementary adversarial supervision leveraging discriminative features from both real and fake score models. Finally, a Preference-Guided Refinement stage further aligns the student with perceptual quality preferences. Extensive experiments demonstrate that DUO-VSR achieves superior visual quality and efficiency over previous one-step VSR approaches.

Problem

• video super-resolution
• diffusion models
• distillation
• training instability
• supervision deficiency

Innovation

• Dual-Stream Distillation
• One-Step Video Super-Resolution
• Distribution Matching Distillation
• RFS-GAN
• Preference-Guided Refinement