Research map / Video / Video segmentation and tracking
STAR: Spatial-Temporal Augmentation with Text-to-Video Models for Real-World Video Super-Resolution
Rui Xie, Yinhong Liu, Penghao Zhou, Chen Zhao, Jun Zhou, Kai Zhang, Zhenyu Zhang, Jian Yang, Zhenheng Yang, Ying Tai
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add LIEM on spatial blocks (ablation) | STAR · UDM10 | PSNR: 23.14 -> 23.61 | Improved |
| add LIEM on temporal blocks (ablation) | STAR · UDM10 | PSNR: 23.14 -> 23.65 | Improved |
| add LIEM on both spatial and temporal blocks (ablation) | STAR · UDM10 | LPIPS: 0.2015 -> 0.1943 | Improved |
| insert LIEM at position (i) instead of (ii) (ablation) | STAR · UDM10 | LPIPS: 0.2363 -> 0.1943 | Improved |
| insert LIEM at position (i) instead of (iii) (ablation) | STAR · UDM10 | Ewarp (x1e-3): 2.74 -> 1.99 | Mixed |
| separate frequencies with direct-type DF loss (ablation) | STAR · UDM10 | LPIPS: 0.1943 -> 0.1903 | Improved |
| use inverse instead of direct weighting of frequencies (ablation) | STAR · UDM10 | Ewarp (x1e-3): 2.69 -> 2.83 | Hurt |
| use alpha 2 instead of 1 in c(t) with linear b(t) (ablation) | STAR · UDM10 | LPIPS: 0.1903 -> 0.1885 | Improved |
| use exponential b(t) instead of linear with alpha 2 (ablation) | STAR · UDM10 | LPIPS: 0.1885 -> 0.1990 | Hurt |
| increase DF loss weight beta from 1.0 to 2.0 (ablation) | STAR · UDM10 | PSNR: 23.91 -> 24.41 | Mixed |
| decrease DF loss weight beta from 1.0 to 0.25 (ablation) | STAR · UDM10 | LPIPS: 0.1885 -> 0.1825 | Mixed |
| scale T2V backbone from CogVideoX-2B to CogVideoX-5B (ablation) | STAR · OpenVid cropped 480x720 | SSIM: 0.7112 -> 0.7400 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Upscale-A-Video — Upscale-A-Video: Temporal-Consistent Diffusion Model for Real-World Video Super-Resolution (CVPR 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Upscale-A-Video: Temporal-Consistent Diffusion Model for Real-World Video Super-Resolution (CVPR 2024 · similarity 0.90)
- Learning Spatial Adaptation and Temporal Coherence in Diffusion Models for Video Super-Resolution (CVPR 2024 · similarity 0.89)
- One-Step Diffusion for Detail-Rich and Temporally Consistent Video Super-Resolution (NeurIPS 2025 · similarity 0.89)
- VideoGigaGAN: Towards Detail-rich Video Super-Resolution (CVPR 2025 · similarity 0.88)
- NSARM: Next-Scale Autoregressive Modeling for Robust Real-World Image Super-Resolution (NeurIPS 2026 · similarity 0.88)
- VideoGuide: Improving Video Diffusion Models without Training Through a Teacher's Guide (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
