DUO-VSR: Dual-Stream Distillation for One-Step Video Super-Resolution
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910066280497152 |
|---|---|
| author | Lv, Zhengyao Xia, Menghan Wang, Xintao Wong, Kwan-Yee K. |
| author_facet | Lv, Zhengyao Xia, Menghan Wang, Xintao Wong, Kwan-Yee K. |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_22271 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DUO-VSR: Dual-Stream Distillation for One-Step Video Super-Resolution Lv, Zhengyao Xia, Menghan Wang, Xintao Wong, Kwan-Yee K. Computer Vision and Pattern Recognition 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. |
| title | DUO-VSR: Dual-Stream Distillation for One-Step Video Super-Resolution |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.22271 |