Research map / Detection and segmentation / Open-set and few-shot detection
Optical Flow With Semantic Segmentation and Localized Layers
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 semantic optical flow with class-specific motion models and localized layers instead of DiscreteFlow | Optical flow estimation · KITTI 2015 test | Fl-all (All px): 22.38% -> 16.81% | Helped |
| Use semantic optical flow instead of DiscreteFlow | Optical flow estimation · KITTI 2015 test | Fl-bg (All px): 21.53% -> 14.63% | Helped |
| Use semantic optical flow instead of DiscreteFlow | Optical flow estimation · KITTI 2015 test | Fl-fg (All px): 26.68% -> 27.73% | Hurt |
| Use semantic optical flow instead of DiscreteFlow | Optical flow estimation · KITTI 2015 test | Fl-all (Nocc): 12.18% -> 10.86% | Helped |
| Use semantic optical flow instead of DiscreteFlow | Optical flow estimation · KITTI 2015 test | Fl-bg (Nocc): 9.96% -> 8.11% | Helped |
| Use semantic optical flow instead of DiscreteFlow | Optical flow estimation · KITTI 2015 test | Fl-fg (Nocc): 22.17% -> 23.28% | Hurt |
| Use semantic optical flow instead of FullFlow | Optical flow estimation · KITTI 2015 test | Fl-all (All px): 24.26% -> 16.81% | Helped |
| Use class-specific motion models for semantic regions | SOF · KITTI 2015 test | Fl-all (All px) | Helped |
| Use localized layered flow around independently moving objects | SOF · Natural YouTube sequences | Optical flow quality | Helped |
| Use flow to refine the semantic segmentation | SOF · Natural YouTube sequences | Segmentation quality | Helped |
| Apply SOF when semantic segmentation is mostly unknown or lacks a relevant class motion model | Optical flow estimation · Middlebury | Accuracy: 0.004 better | No clear effect |
Closest papers by meaning
- Semantic Flow: Learning Semantic Fields of Dynamic Scenes from Monocular Videos (ICLR 2024 · similarity 0.89)
- UnSAMFlow: Unsupervised Optical Flow Guided by Segment Anything Model (CVPR 2024 · similarity 0.88)
- FlowFeat: Pixel-Dense Embedding of Motion Profiles (NeurIPS 2025 · similarity 0.87)
- Learning Segmentation from Point Trajectories (NeurIPS 2024 · similarity 0.87)
- SemFlow: Binding Semantic Segmentation and Image Synthesis via Rectified Flow (NeurIPS 2024 · similarity 0.86)
- Segment Any Motion in Videos (CVPR 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.