Research map / Video / Video segmentation and tracking
Learning Degradation-Robust Spatiotemporal Frequency-Transformer for Video Super-Resolution
Zhongwei Qiu, Huan Yang, Jianlong Fu, Daochang Liu, Chang Xu, Dongmei Fu
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 |
|---|---|---|---|
| learn in frequency domain with CNN instead of pixel domain (ablation) | BasicVSR · REDS4 compressed CRF25 | PSNR: 25.93 -> 26.87 | Improved |
| use Transformer instead of CNN in frequency domain (ablation) | Frequency + CNN · REDS4 compressed CRF25 | PSNR: 26.87 -> 27.15 | Improved |
| add frequency attention in place of vanilla attention (ablation) | Frequency + Transformer · REDS4 compressed CRF25 | PSNR: 27.15 -> 27.28 | Improved |
| add dual frequency attention instead of local frequency attention (ablation) | FTVSR · REDS CRF25 | PSNR: 27.23 -> 27.53 | Improved |
| add dual frequency attention instead of global frequency attention (ablation) | FTVSR · REDS blur downsampling (BD) | PSNR: 35.87 -> 36.33 | Improved |
| add dual frequency attention instead of local frequency attention (ablation) | FTVSR · REDS random Gaussian noise 15 | PSNR: 26.05 -> 26.35 | Improved |
| use global instead of local frequency attention (ablation) | FTVSR · REDS CRF25 | PSNR: 27.23 -> 27.14 | Hurt |
| add space-then-time divided frequency attention (Lambda_ST) to base attention (ablation) | Frequency + Transformer · REDS CRF35 | PSNR: 24.03 -> 24.22 | Improved |
| add time-only frequency attention (Lambda_T) to base attention (ablation) | Frequency + Transformer · REDS CRF25 | PSNR: 27.15 -> 27.11 | No clear effect |
| order space-frequency before time-frequency attention (Lambda_ST vs Lambda_TS) (ablation) | FTVSR · REDS CRF25 | PSNR: 27.24 -> 27.28 | Improved |
| replace TTVSR with FTVSR | TTVSR · REDS4 uncompressed 4x | PSNR: 32.12 -> 32.42 | Improved |
| replace COMISR with FTVSR on compressed videos | COMISR · REDS4 compressed CRF25 | PSNR: 26.47 -> 28.13 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- BasicVSR++ — BasicVSR: The Search for Essential Components in Video Super-Resolution and Beyond (CVPR 2021) · better on 4 of 6 shared results
- COMISR — COMISR: Compression-Informed Video Super-Resolution (ICCV 2021) · better on 6 of 6 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
- Learning Spatiotemporal Frequency-Transformer for Compressed Video Super-Resolution (ECCV 2022 · similarity 0.96)
- Learning Trajectory-Aware Transformer for Video Super-Resolution (CVPR 2022 · similarity 0.89)
- Towards Interpretable Video Super-Resolution via Alternating Optimization (ECCV 2022 · similarity 0.88)
- FMA-Net: Flow-Guided Dynamic Filtering and Iterative Feature Refinement with Multi-Attention for Joint Video Super-Resolution and Deblurring (CVPR 2024 · similarity 0.87)
- Self-supervised ControlNet with Spatio-Temporal Mamba for Real-world Video Super-resolution (CVPR 2025 · similarity 0.87)
- Hazy Low-Quality Satellite Video Restoration Via Learning Optimal Joint Degradation Patterns and Continuous-Scale Super-Resolution Reconstruction (CVPR 2025 · similarity 0.86)
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.
