Research map / Video / Video segmentation and tracking
3D Rigid Motion Segmentation with Mixed and Unknown Number of Models
Xun Xu, Loong‐Fah Cheong, Zhuwen Li
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 |
|---|---|---|---|
| fuse affine, homography and fundamental affinities via subset constrained clustering (ablation) | single-model motion segmentation · Hopkins155 | classification error (%): 0.59 -> 0.31 | Improved |
| fuse models with co-regularization instead of single homography (ablation) | homography model · Hopkins155 | classification error (%): 0.71 -> 0.46 | Improved |
| fuse models with subset constrained clustering (ablation) | best single model · KT3DMoSeg | mean classification error (%): 11.45 -> 8.08 | Improved |
| fuse models with co-regularization (ablation) | best single model · KT3DMoSeg | mean classification error (%): 11.45 -> 7.92 | Improved |
| fuse models with subset constrained clustering (ablation) | affine model · MTPV62 | classification error (%): 0.82 -> 0.65 | Improved |
| use fundamental matrix instead of affine as single model (ablation) | single-model motion segmentation · Hopkins155 | classification error (%): 0.59 -> 1.79 | Hurt |
| use affine model instead of homography on strong perspective data (ablation) | single-model motion segmentation · KT3DMoSeg | mean classification error (%): 11.45 -> 15.76 | Hurt |
| use homography model for model selection with 2 to Mmax candidates over 1 to Mmax (ablation) | NCRE model selection · Hopkins155 | correct rate (%): 87.10 -> 92.26 | Improved |
| replace MSSC with NCRE homography model selection | MSSC · Hopkins155 | correct rate (%): 85.81 -> 92.26 | Improved |
| fuse three models for model selection instead of single homography (ablation) | NCRE model selection · Hopkins155 | classification error (%): 3.95 -> 3.17 | Improved |
| fuse three models for model selection instead of single homography (ablation) | NCRE model selection · Hopkins155 | correct rate | No clear effect |
| replace MSSC with NCRE affine model | MSSC · MTPV62 | classification error (%): 5.09 -> 2.78 | Improved |
Closest papers by meaning
- Motion Segmentation by Exploiting Complementary Geometric Models (CVPR 2018 · similarity 0.96)
- What to Select: Pursuing Consistent Motion Segmentation from Multiple Geometric Models (AAAI 2021 · similarity 0.91)
- Multi-frame Motion Segmentation by Combining Two-Frame Results (IJCV 2022 · similarity 0.88)
- It’s Moving! A Probabilistic Model for Causal Motion Segmentation in Moving Camera Videos (ECCV 2016 · similarity 0.87)
- Segmenting the Motion Components of a Video: A Long-Term Unsupervised Model (TPAMI 2025 · similarity 0.87)
- Learning Segmentation from Point Trajectories (NeurIPS 2024 · similarity 0.87)
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.
