Research map / Graph learning / Graph self-supervised learning
PlantPose: Universal Plant Skeleton Estimation via Tree-constrained Graph Generation
Xinpeng Liu, Hiroaki Santo, Yosuke Toda, Fumio Okura
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 training-time tree constraint (SFS layer) over test-time MST only (ablation) | PlantPose · Grapevine (domain-specific training) | SMD: 1.47 x 10^-4 -> 1.03 x 10^-4 | Improved |
| add training-time tree constraint over test-time MST only (ablation) | PlantPose · Synthetic (domain-specific training) | TOPO F1: 0.965 -> 0.977 | Improved |
| add training-time tree constraint over test-time MST only (ablation) | PlantPose · Root (domain-specific training) | TOPO F1: 0.799 -> 0.833 | Improved |
| add tree constraint (test-time MST) to unconstrained generator (ablation) | RelationFormer · Synthetic (domain-specific training) | tree rate: 36.2 -> 100.0 | Improved |
| add test-time MST to unconstrained generator (ablation) | RelationFormer · Grapevine (domain-specific training) | TOPO F1: 0.708 -> 0.867 | Improved |
| add training-time constraint over test-time constraint | PlantPose · Root (out-of-domain) | SMD: 6.06 x 10^-2 -> 2.70 x 10^-2 | Improved |
| add training-time constraint over test-time constraint | PlantPose · Root (out-of-domain) | TOPO F1: 0.592 -> 0.555 | Hurt |
| add training-time constraint over test-time constraint | PlantPose · Web-sourced (generalized training) | TOPO F1: 0.606 -> 0.594 | Hurt |
| add training-time constraint over test-time constraint | PlantPose · Self-captured (generalized training) | SMD: 1.68 x 10^-3 -> 1.92 x 10^-3 | Hurt |
| add training-time constraint over test-time constraint | PlantPose · Tree synthetic (out-of-domain) | TOPO F1: 0.478 -> 0.483 | No clear effect |
| add training-time constraint over test-time constraint | PlantPose · LVD2021 vein mask (generalized training) | TOPO F1: 0.862 -> 0.888 | Improved |
| add training-time constraint over test-time constraint | PlantPose · Thickened L-system (out-of-domain) | TOPO F1: 0.796 -> 0.841 | Improved |
Closest papers by meaning
- Learning to Infer Parameterized Representations of Plants from 3D Scans (CVPR 2026 · similarity 0.88)
- Tree Skeletonization from 3D Point Clouds by Denoising Diffusion (ICCV 2025 · similarity 0.88)
- Neural Hierarchical Decomposition for Single Image Plant Modeling (CVPR 2025 · similarity 0.88)
- Demeter: A Parametric Model of Crop Plant Morphology from the Real World (ICCV 2025 · similarity 0.88)
- SVDTree: Semantic Voxel Diffusion for Single Image Tree Reconstruction (CVPR 2024 · similarity 0.86)
- BioCLIP: A Vision Foundation Model for the Tree of Life (CVPR 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.
