Research map / Video / Video segmentation and tracking
Rethinking Alignment in Video Super-Resolution Transformers
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace pixel-level alignment with feature-level Patch Alignment using nearest-neighbor resampling (ablation) | VSR Transformer · REDS4 | PSNR: 31.00 -> 31.17 | Helped |
| Use bilinear instead of nearest-neighbor resampling for feature-level Patch Alignment (ablation) | VSR Transformer · REDS4 | PSNR: 31.17 -> 31.00 | Hurt |
| Add image-level Patch Alignment with nearest-neighbor resampling (ablation) | VSR Transformer · REDS4 | PSNR: 30.92 -> 31.11 | Helped |
| Add feature alignment with bilinear resampling (ablation) | VSR Transformer · REDS4 | PSNR: 30.92 -> 31.06 | Helped |
| Use nearest-neighbor instead of bilinear resampling for feature alignment (ablation) | VSR Transformer · REDS4 | PSNR: 31.06 -> 31.11 | Helped |
| Add image alignment with bilinear resampling (ablation) | VSR Transformer · REDS4 | PSNR: 30.92 -> 30.84 | Hurt |
| Remove alignment and use window size 16 rather than 8 (ablation) | VSR Transformer · REDS4 | PSNR: 30.56 -> 30.81 | Helped |
| Remove alignment from a VSR Transformer (ablation) | VSR Transformer · REDS4 | PSNR: 30.43 -> 30.56 | Helped |
| Add Patch Alignment (ablation) | PSRT-recurrent · Vimeo-90K-T | PSNR: 37.87 -> 37.80 | Hurt |
| Add Patch Alignment (ablation) | PSRT-recurrent · Vid4 | PSNR: 27.71 -> 27.72 | No clear effect |
| Replace BasicVSR++ CNN backbone with Transformer blocks (ablation) | BasicVSR++ recurrent framework · REDS4 | PSNR: 31.38 -> 31.88 | Helped |
| Replace FGDC with Patch Alignment in the recurrent Transformer (ablation) | Recurrent VSR Transformer · REDS4 | PSNR: 31.89 -> 31.88 | No clear effect |
Closest papers by meaning
- Sequence Matters: Harnessing Video Models in 3D Super-Resolution (AAAI 2025 · similarity 0.89)
- PatchVSR: Breaking Video Diffusion Resolution Limits with Patch-wise Video Super-Resolution (CVPR 2025 · similarity 0.88)
- Enhancing Video Super-Resolution via Implicit Resampling-based Alignment (CVPR 2024 · similarity 0.87)
- VideoGigaGAN: Towards Detail-rich Video Super-Resolution (CVPR 2025 · similarity 0.86)
- Semantic Lens: Instance-Centric Semantic Alignment for Video Super-resolution (AAAI 2024 · similarity 0.86)
- SeeClear: Semantic Distillation Enhances Pixel Condensation for Video Super-Resolution (NeurIPS 2024 · 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.