Research map / Image generation / Image restoration
Learning to Synthesize Motion Blur
Check 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 the proposed line-prediction model instead of the strongest reported baseline, Super SloMo | motion blur synthesis from image pairs · Real test dataset | PSNR: 33.64 ± 4.66 -> 34.14 ± 4.65 | Helped |
| use the proposed line-prediction model instead of the strongest reported baseline, Super SloMo | motion blur synthesis from image pairs · Real test dataset | SSIM: 0.958 ± 0.048 -> 0.963 ± 0.045 | Helped |
| use the proposed model rather than Super SloMo | motion blur synthesis from image pairs · Real test dataset | Runtime (ms): 13.7 × 10^2 -> 43.7 | Helped |
| add learned sample weights rather than use uniform weights (ablation) | line-prediction motion blur model · Real test dataset | PSNR: 33.88 ± 4.68 -> 34.14 ± 4.65 | Helped |
| add learned sample weights rather than use uniform weights (ablation) | line-prediction motion blur model · Real test dataset | SSIM: 0.959 ± 0.050 -> 0.963 ± 0.045 | Helped |
| add line prediction rather than directly predict the blurred image (ablation) | motion blur synthesis model · Real test dataset | PSNR: 33.97 ± 4.53 -> 34.14 ± 4.65 | Helped |
| add line prediction rather than use separable kernel prediction (ablation) | motion blur synthesis model · Real test dataset | PSNR: 33.73 ± 4.31 -> 34.14 ± 4.65 | Helped |
| add line prediction rather than use separable kernel prediction (ablation) | motion blur synthesis model · Real test dataset | SSIM: 0.961 ± 0.045 -> 0.963 ± 0.045 | Mixed |
| pretrain the line-prediction model using optical-flow training data (ablation) | motion blur synthesis model · Training | Performance | No clear effect |
| use negative backward flow rather than forward flow to produce the blur | optical-flow baselines · Real test dataset | PSNR: ∼5 | Helped |
Closest papers by meaning
- From Motion Blur to Motion Flow: A Deep Learning Solution for Removing Heterogeneous Motion Blur (CVPR 2017 · similarity 0.90)
- Bringing Alive Blurred Moments (CVPR 2019 · similarity 0.90)
- Parameterized Blur Kernel Prior Learning for Local Motion Deblurring (CVPR 2025 · similarity 0.87)
- Dynamic Gaussian Splatting from Defocused and Motion-blurred Monocular Videos (NeurIPS 2025 · similarity 0.86)
- BlurDM: A Blur Diffusion Model for Image Deblurring (NeurIPS 2025 · similarity 0.86)
- Efficient Multi-scale Network with Learnable Discrete Wavelet Transform for Blind Motion Deblurring (CVPR 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.