Research map / 3D vision / Depth and stereo
Exploiting Semantic Information and Deep Matching for Optical Flow
Min Bai, Wenjie Luo, Kaustav Kundu, Raquel Urtasun
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 |
|---|---|---|---|
| add 3D detection + CAD fitting to instance segmentation A (ablation) | object-aware flow · KITTI 2015 flow (train, 5-fold) | Fl-fg error (all px, %): 28.11 -> 25.08 | Improved |
| add 3D detection + CAD fitting to instance segmentation B (ablation) | object-aware flow · KITTI 2015 flow (train, 5-fold) | Fl-fg error (all px, %): 27.31 -> 24.79 | Improved |
| estimate F from confident matching-network matches instead of DiscreteFlow (ablation) | object-aware flow · KITTI 2015 flow (train, 5-fold) | foreground error (all px, %): 26.05 -> 24.79 | Improved |
| estimate F from confident matching-network matches instead of EpicFlow (ablation) | object-aware flow · KITTI 2015 flow (train, 5-fold) | foreground error (all px, %): 27.58 -> 24.79 | Improved |
| use perfect (ground truth) object masks instead of detected masks (ablation) | object-aware flow · KITTI 2015 flow (train, 5-fold) | foreground error (all px, %): 19.72 -> 18.62 | Improved |
| use object-aware epipolar flow instead of SOF | KITTI test · KITTI 2015 flow (test) | Fl-all error (all px, %): 16.81 -> 11.62 | Improved |
| use object-aware epipolar flow instead of DiscreteFlow (background) | KITTI test · KITTI 2015 flow (test) | Fl-bg error (all px, %): 21.53 -> 8.61 | Improved |
| use object-aware epipolar flow instead of DiscreteFlow (foreground) | KITTI test · KITTI 2015 flow (test) | Fl-fg error (all px, %): 26.68 -> 26.69 | No clear effect |
| use object-aware epipolar flow instead of MotionSLIC (foreground) | KITTI test · KITTI 2015 flow (test) | Fl-fg error (all px, %): 66.21 -> 26.69 | Improved |
Closest papers by meaning
- SemARFlow: Injecting Semantics into Unsupervised Optical Flow Estimation for Autonomous Driving (ICCV 2023 · similarity 0.91)
- Mono-SF: Multi-View Geometry Meets Single-View Depth for Monocular Scene Flow Estimation of Dynamic Traffic Scenes (ICCV 2019 · similarity 0.89)
- Just Go With the Flow: Self-Supervised Scene Flow Estimation (CVPR 2020 · similarity 0.89)
- M2Flow: A Motion Information Fusion Framework for Enhanced Unsupervised Optical Flow Estimation in Autonomous Driving (AAAI 2025 · similarity 0.89)
- ICP-Flow: LiDAR Scene Flow Estimation with ICP (CVPR 2024 · similarity 0.87)
- UnSAMFlow: Unsupervised Optical Flow Guided by Segment Anything Model (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.
