Research map / 3D vision / Depth and stereo
Optical Flow Estimation from a Single Motion-blurred Image
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 |
|---|---|---|---|
| Fine-tune the pretrained model on GoPro blur data | B2F-net · GoPro blur test | EPE: 2.077 -> 2.038 | Helped |
| Pretrain on synthetic Monkaa data, then fine-tune on NfS blur data | B2F-net · NfS blur test | EPE: 1.958 | Helped |
| Estimate flow directly from blurred input rather than from restored endpoint frames | B2F-net · GoPro blur test | EPE: 5.875 -> 2.077 | Helped |
| Remove spatial transformer networks (STN) (ablation) | B2F-net · GoPro blur test | EPE: 2.077 -> 2.263 | Hurt |
| Remove feature refining blocks (RB) (ablation) | B2F-net · GoPro blur test | EPE: 2.077 -> 2.383 | Hurt |
| Remove both spatial transformer networks and feature refining blocks (ablation) | B2F-net · GoPro blur test | EPE: 2.077 -> 2.748 | Hurt |
| Use a U-net-like flow architecture without explicit feature decoding (ablation) | B2F-net · GoPro blur test | EPE: 2.077 -> 2.748 | Hurt |
| Add an L1 photometric loss for supervision (ablation) | B2F-net · Motion-blurred flow estimation | EPE | No clear effect |
| Add a smoothness-loss regularizer (ablation) | B2F-net · Motion-blurred flow estimation | EPE | No clear effect |
Closest papers by meaning
- Parametric Object Motion From Blur (CVPR 2016 · similarity 0.92)
- From Motion Blur to Motion Flow: A Deep Learning Solution for Removing Heterogeneous Motion Blur (CVPR 2017 · similarity 0.92)
- Motion-adaptive Separable Collaborative Filters for Blind Motion Deblurring (CVPR 2024 · similarity 0.88)
- Recurrent Partial Kernel Network for Efficient Optical Flow Estimation (AAAI 2024 · similarity 0.87)
- Taming generative video models for zero-shot optical flow extraction (NeurIPS 2025 · similarity 0.86)
- FlowFeat: Pixel-Dense Embedding of Motion Profiles (NeurIPS 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.