Research map / Image generation / Image restoration
Deep Homography Mixture for Single Image Rolling Shutter Correction
Weilong Yan, Robby T. Tan, Bing Zeng, Shuaicheng Liu
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 |
|---|---|---|---|
| use Gaussian weight map instead of average weight map (ablation) | homography mixture · RS-Homo (ablation, 5 categories) | PSNR: 23.74 -> 26.15 | Improved |
| use Gaussian weight map instead of single weight map (ablation) | homography mixture · RS-Homo (ablation, 5 categories) | EPE: 5.18 -> 4.10 | Improved |
| use 8 blocks instead of 4 (ablation) | homography mixture · RS-Homo (ablation, 5 categories) | EPE: 4.25 -> 4.10 | Improved |
| use 8 blocks instead of 12 (ablation) | homography mixture · RS-Homo (ablation, 5 categories) | PSNR: 25.61 -> 26.15 | Improved |
| predict basis coefficients instead of corner offsets (ablation) | homography mixture · RS-Homo (ablation, 5 categories) | EPE: 15.58 -> 4.10 | Improved |
| learn bases from natural videos with PCA instead of fixed 8 bases (ablation) | homography mixture · RS-Homo (ablation, 5 categories) | EPE: 4.27 -> 4.10 | Improved |
| use 12 learned bases instead of 8 (ablation) | homography mixture · RS-Homo (ablation, 5 categories) | PSNR: 26.15 -> 25.88 | No clear effect |
| use our method versus Kandula et al. on building dataset | single-image RS correction · Building dataset | EPE: 2.84 -> 1.25 | Improved |
| use our method versus previous single-frame method requiring depth | single-image RS correction · Carla-RS | PSNR: 18.47 -> 20.62 | Improved |
| use our single-frame method versus multi-frame SUNet | single-image RS correction · Carla-RS | PSNR: 29.18 -> 20.62 | Hurt |
Closest papers by meaning
- Learning Structure-And-Motion-Aware Rolling Shutter Correction (CVPR 2019 · similarity 0.90)
- Rolling Shutter Homography and its Applications (TPAMI 2020 · similarity 0.90)
- Single Image Rolling Shutter Removal with Diffusion Models (AAAI 2025 · similarity 0.90)
- Unrolling the Shutter: CNN to Correct Motion Distortions (CVPR 2017 · similarity 0.89)
- Neural Global Shutter: Learn To Restore Video From a Rolling Shutter Camera With Global Reset Feature (CVPR 2022 · similarity 0.89)
- Rolling Shutter Correction with Intermediate Distortion Flow Estimation (CVPR 2024 · similarity 0.88)
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.
