Research map / 3D vision / 3D generation
AutoGPart: Intermediate Supervision Search for Generalizable 3D Part Segmentation
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 |
|---|---|---|---|
| Add automatically searched intermediate supervision | PointNet++ · Mobility-based segmentation OOD test | Segmentation MIoU: 61.0 -> 66.5 | Helped |
| Add automatically searched intermediate supervision | DGCNN · Mobility-based segmentation OOD test | Segmentation MIoU: 68.1 -> 73.8 | Helped |
| Add automatically searched intermediate supervision | PointNet++ · Primitive fitting OOD test | MIoU: 71.6 -> 76.5 | Helped |
| Add automatically searched intermediate supervision | DGCNN · Primitive fitting OOD test | MIoU: 68.0 -> 73.4 | Helped |
| Add automatically searched intermediate supervision to HPNet's first learning stage | HPNet · Primitive fitting OOD test | MIoU: 79.5 -> 80.4 | Helped |
| Add automatically searched intermediate supervision | PointNet++ · Semantic-based segmentation OOD test | Mean Recall: 32.0 -> 33.9 | Helped |
| Use cross-validation rather than a single train-validation split to estimate supervision reward (ablation) | AutoGPart PointNet++ · Mobility-based segmentation OOD test | Segmentation MIoU: 64.6 -> 66.5 | Helped |
| Use generalization gap rather than average validation performance as the reward (ablation) | AutoGPart PointNet++ · Mobility-based segmentation OOD test | Segmentation MIoU: 64.9 -> 66.5 | Helped |
| Retain the larger part-aware feature and operator space (ablation) | AutoGPart PointNet++ · Mobility-based segmentation OOD test | Segmentation MIoU: 63.1 / 64.1 / 65.1 / 63.5 -> 66.5 | Helped |
| Retain the larger part-aware feature and operator space (ablation) | AutoGPart DGCNN · Mobility-based segmentation OOD test | Segmentation MIoU: 70.6 / 69.5 / 71.6 / 71.8 -> 73.8 | Helped |
| Use greedy supervision selection rather than probability-ranked Top1, Top2, or Top3 selection (ablation) | AutoGPart PointNet++ · Mobility-based segmentation OOD test | Segmentation MIoU: 65.6 / 64.5 / 62.8 -> 66.5 | Helped |
| Use greedy supervision selection rather than probability-ranked Top1, Top2, or Top3 selection (ablation) | AutoGPart DGCNN · Mobility-based segmentation OOD test | Segmentation MIoU: 70.1 / 67.8 / 71.6 -> 73.8 | Helped |
Closest papers by meaning
- PartSAM: A Scalable Promptable Part Segmentation Model Trained on Native 3D Data (ICLR 2026 · similarity 0.88)
- GeoSAM2: Unleashing the Power of SAM2 for 3D Part Segmentation (CVPR 2026 · similarity 0.87)
- SegGraph: Leveraging Graphs of SAM Segments for Few-Shot 3D Part Segmentation (NeurIPS 2025 · similarity 0.87)
- PartDistillation: Learning Parts From Instance Segmentation (CVPR 2023 · similarity 0.87)
- PartNeXt: A Next-Generation Dataset for Fine-Grained and Hierarchical 3D Part Understanding (NeurIPS 2025 · similarity 0.85)
- G-FARS: Gradient-Field-based Auto-Regressive Sampling for 3D Part Grouping (CVPR 2024 · 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.