Research map / Video / Video segmentation and tracking
Seeing Tree Structure from Vibration
Tianfan Xue, Jiajun Wu, Zhoutong Zhang, Chengkai Zhang, Joshua B. Tenenbaum, William T. Freeman
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 vibration spectrum cues to appearance cues (ablation) | tree structure inference · Tree structure videos (avg over 17 videos) | parent accuracy (%): 65.7 -> 79.3 | Improved |
| add vibration spectrum cues to appearance cues (ablation) | tree structure inference · Tree structure videos (avg over 17 videos) | edit distance: 13.7 -> 7.0 | Improved |
| use spectral motion features instead of raw optical flow (ablation) | appearance + motion · Tree structure videos (avg over 17 videos) | parent accuracy (%): 73.5 -> 79.3 | Improved |
| use spectral motion features instead of KLT tracking (ablation) | appearance + motion · Tree structure videos (avg over 17 videos) | parent accuracy (%): 73.8 -> 79.3 | Improved |
| use spectral motion features instead of raw optical flow (ablation) | appearance + motion · Tree structure videos (avg over 17 videos) | edit distance: 9.5 -> 7.0 | Improved |
| use appearance plus vibration instead of hierarchical motion segmentation | tree structure inference · Tree structure videos (avg over 17 videos) | parent accuracy (%): 56.3 -> 79.3 | Improved |
| add vibration cues on tree video A2 (ablation) | appearance only · Tree video A2 | parent accuracy (%): 31 -> 45 | Improved |
| add vibration cues on tree video A3 (ablation) | appearance only · Tree video A3 | parent accuracy (%): 90 -> 100 | Improved |
| use ours on retinal vessels instead of A+Flow or A+Track | structure inference · Retinal video | edit distance: 7 -> 4 | Improved |
| increase video frame rate from 30 to 200 FPS | Ours · High-speed videos H1-H6 | parent accuracy (%): 69 -> 77 | Mixed |
Closest papers by meaning
- Reconstructing Evolving Tree Structures in Time Lapse Sequences by Enforcing Time-Consistency (TPAMI 2017 · similarity 0.84)
- Unsupervised Discovery of Parts, Structure, and Dynamics (ICLR 2019 · similarity 0.84)
- Seeing the Wind from a Falling Leaf (NeurIPS 2025 · similarity 0.83)
- SVDTree: Semantic Voxel Diffusion for Single Image Tree Reconstruction (CVPR 2024 · similarity 0.83)
- Generative Perception of Shape and Material from Differential Motion (NeurIPS 2025 · similarity 0.82)
- COCOTree: A Dataset and Benchmark for Open Tree-Structured Visual Decomposition (NeurIPS 2026 · similarity 0.82)
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.
