Research map / Video / Video segmentation and tracking
Frame-Recurrent Video Super-Resolution
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 |
|---|---|---|---|
| Use recurrent HR-frame propagation instead of sliding-window multi-frame estimation | FRVSR versus VSR · Vid4 | PSNR: 26.25 -> 26.69 | Helped |
| Use recurrent HR-frame propagation instead of single-image super-resolution | FRVSR versus SISR · Vid4 | PSNR: 24.96 -> 26.69 | Helped |
| Use recurrent HR-frame propagation instead of sliding-window multi-frame estimation | FRVSR versus VSR · Vid4 | SSIM: 0.803 -> 0.822 | Helped |
| Use recurrent HR-frame propagation instead of single-image super-resolution | FRVSR versus SISR · Vid4 | SSIM: 0.721 -> 0.822 | Helped |
| Use FRVSR rather than VSR on degraded inputs | FRVSR · YT10, Gaussian noise σ = 0.025 | PSNR: 30.36 -> 30.84 | Helped |
| Use FRVSR rather than VSR on degraded inputs | FRVSR · YT10, Gaussian noise σ = 0.075 | PSNR: 28.42 -> 28.62 | Helped |
| Use FRVSR rather than VSR on degraded inputs | FRVSR · YT10, JPG 40 | PSNR: 28.12 -> 28.30 | Helped |
| Use FRVSR rather than VSR on degraded inputs | FRVSR · YT10, JPG 70 | PSNR: 29.07 -> 29.29 | Helped |
| Increase training clip length from 2 to 5 frames (ablation) | FRVSR · YT10 | video PSNR: 31.60 -> 32.01 | Helped |
| Increase training clip length from 5 to 10 frames (ablation) | FRVSR · YT10 | video PSNR: 32.01 -> 32.10 | Helped |
| Increase training clip length beyond 5 frames (ablation) | FRVSR · YT10 | video PSNR: 32.01 -> 32.10 | No clear effect |
| Train/evaluate with larger Gaussian blur sizes rather than the preferred σ = 1.5 (ablation) | FRVSR · YT10 | PSNR | Hurt |
Closest papers by meaning
- Recurrent Back-Projection Network for Video Super-Resolution (CVPR 2019 · similarity 0.93)
- Frame and Feature-Context Video Super-Resolution (AAAI 2019 · similarity 0.93)
- Video Face Super-Resolution with Motion-Adaptive Feedback Cell (AAAI 2020 · similarity 0.92)
- Image Super-Resolution via Dual-State Recurrent Networks (CVPR 2018 · similarity 0.91)
- Real-Time Video Super-Resolution With Spatio-Temporal Networks and Motion Compensation (CVPR 2017 · similarity 0.91)
- SAVSR: Arbitrary-Scale Video Super-Resolution via a Learned Scale-Adaptive Network (AAAI 2024 · similarity 0.89)
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.