Research map / 3D vision / Depth and stereo
Fast Two-View Motion Segmentation Using Christoffel Polynomials
Bengisu Ozbay, Octavia Camps, Mario Sznaier
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 |
|---|---|---|---|
| raise the outlier ratio from 0.9 to 1.5 (ablation) | Christoffel motion segmentation · Hopkins with outliers | segmentation error (%): 7.79 -> 9.37 | Hurt |
| raise the outlier ratio from 0.9 to 2.1 (ablation) | Christoffel motion segmentation · Hopkins with outliers | segmentation error (%): 7.79 -> 10.72 | Hurt |
| use frames 1 and 10 instead of 1st and last frame (smaller displacement) (ablation) | Christoffel motion segmentation · Hopkins-clean | segmentation error (%): 7.4 -> 9.9 | Hurt |
| use frames 1 and 10 instead of 1st and last frame (smaller displacement) (ablation) | RPA · Hopkins-clean | segmentation error (%): 6.4 -> 5.5 | Improved |
| replace RPA with the proposed method | segmentation pipeline · BCD (3 motions) | processing time per scene (s): 210 -> 0.09 | Improved |
| replace RPA with the proposed method | segmentation pipeline · Adelaide-F | segmentation error (%): 5.49 -> 4.15 | Improved |
| replace CBS with the proposed method | segmentation pipeline · Adelaide-F | processing time per scene (s): 0.69 -> 0.02 | Improved |
| replace RPA with the proposed method | segmentation pipeline · KT3D | segmentation error (%): 46.4 -> 17.2 | Improved |
| replace P-x with the proposed method | segmentation pipeline · KT3D | segmentation error (%): 15.0 -> 17.2 | Hurt |
| replace P-xp with the proposed method | segmentation pipeline · Hopkins-clean | segmentation error (%): 32.2 -> 7.42 | Improved |
| replace RPA with the proposed method | segmentation pipeline · Pairwise | segmentation error (%): 2.63 -> 4.93 | Hurt |
| go from two to three motions per scene | Christoffel motion segmentation · Hopkins-clean | segmentation error (%) | Hurt |
Closest papers by meaning
- Multi-frame Motion Segmentation by Combining Two-Frame Results (IJCV 2022 · similarity 0.87)
- Efficient Solution of Point-Line Absolute Pose (CVPR 2024 · similarity 0.85)
- Robust Motion Segmentation From Pairwise Matches (ICCV 2019 · similarity 0.85)
- Learning To Segment Rigid Motions From Two Frames (CVPR 2021 · similarity 0.85)
- Realigning 2D and 3D Object Fragments without Correspondences (TPAMI 2016 · similarity 0.85)
- Learning Segmentation from Point Trajectories (NeurIPS 2024 · similarity 0.84)
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.
