Research map / 3D vision / Depth and stereo
GeoMotion: Rethinking Motion Segmentation via Latent 4D Geometry
Xiankang He, Peile Lin, Ying Cui, Dongyan Guo, Chunhua Shen, Xiaoqin Zhang
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 camera pose features (ablation) | GeoMotion baseline · DAVIS2017 | J&F: 67.9 -> 74.2 | Improved |
| add optical flow features (ablation) | GeoMotion baseline · DAVIS2017 | J&F: 67.9 -> 74.7 | Improved |
| add shallow-layer features (ablation) | GeoMotion baseline · DAVIS2017 | J&F: 67.9 -> 72.4 | Improved |
| combine camera, flow and shallow features (ablation) | GeoMotion · DAVIS2017 | J&F: 80.2 -> 81.4 | Improved |
| add Dynamic Replica to HOI4D training data (ablation) | GeoMotion · DAVIS2017 | J: 65.4 -> 73.8 | Improved |
| scale training data from 1 to 5 datasets (ablation) | GeoMotion · DAVIS2016-Moving | J: 67.9 -> 83.5 | Improved |
| initialize motion decoder from pi3 confidence decoder instead of random (ablation) | GeoMotion · DAVIS (training curve) | IoU J | Improved |
| feed-forward 4D-geometry design vs iterative-optimization methods | GeoMotion · DAVIS2016 | J&F: 84.7 vs 90.9 (SegAnyMotion) | Mixed |
Later papers that compared with this method
1 later accepted paper reports GeoMotion as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- ProMo3D: Probing Motion Cues in Frozen 3D Foundation Models (NeurIPS 2026 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.88)
- Layered Motion Fusion: Lifting Motion Segmentation to 3D in Egocentric Videos (CVPR 2025 · similarity 0.88)
- Segment Any Motion in Videos (CVPR 2025 · similarity 0.87)
- Learning Segmentation from Point Trajectories (NeurIPS 2024 · similarity 0.87)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.86)
- IGGT4D: Streaming 4D Instance-Grounded Geometry Transformer (NeurIPS 2026 · 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.
