Research map / Video / Video segmentation and tracking
Discovering the Physical Parts of an Articulated Object Class From Multiple Videos
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 |
|---|---|---|---|
| Learn the part location model jointly across videos rather than independently per video | Part discovery on tigerL · tigerL, per-video overlap | average part overlap: 0.244 -> 0.296 | Helped |
| Learn the part location model jointly across videos rather than independently per video | Part discovery on tigerR · tigerR, per-video overlap | average part overlap: 0.247 -> 0.274 | Helped |
| Learn the part location model jointly across videos rather than independently per video | Part discovery on horseL · horseL, per-video overlap | average part overlap: 0.267 -> 0.268 | No clear effect |
| Learn the part location model jointly across videos rather than independently per video | Part discovery on horseR · horseR, per-video overlap | average part overlap: 0.239 -> 0.264 | Helped |
| Add the appearance potential (ablation) | Part discovery using the location potential · Across-video overlap | average part overlap: 0.204 -> 0.223 | Helped |
| Add the foreground potential (ablation) | Part discovery using location and appearance potentials · Across-video overlap | average part overlap: 0.223 -> 0.228 | Helped |
| Add the temporal smoothness potential (ablation) | Part discovery with unary potentials · Across-video overlap | average part overlap: 0.228 -> 0.230 | Helped |
| Add the motion potential (ablation) | Part discovery with unary and temporal potentials · Across-video overlap | average part overlap: 0.230 -> 0.243 | Helped |
| Add the edge potential (ablation) | Part discovery with unary, temporal, and motion potentials · Across-video overlap | average part overlap: 0.243 -> 0.245 | Helped |
| Set the number of parts below 10 (ablation) | Full method · Tiger and horse videos | part segmentation quality | Mixed |
| Set the number of parts above 10 (ablation) | Full method · Tiger and horse videos | part segmentation quality | Hurt |
Closest papers by meaning
- Articulation in Motion: Prior-free Part Mobility Analysis for Articulated Objects By Dynamic-Static Disentanglement (ICLR 2026 · similarity 0.89)
- Articulation in Prime: Primitive-Based Articulated Object Understanding from a Single Casual Video (NeurIPS 2026 · similarity 0.88)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.86)
- Articulate your NeRF: Unsupervised articulated object modeling via conditional view synthesis (NeurIPS 2024 · similarity 0.86)
- SAMoR: Motion Modelling for Articulated Objects of Any Skeleton and Topology (NeurIPS 2026 · similarity 0.85)
- Building Interactable Replicas of Complex Articulated Objects via Gaussian Splatting (ICLR 2025 · similarity 0.85)
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.