Research map / 3D vision / 3D generation
PartSTAD: 2D-to-3D Part Segmentation Task Adaptation
Hyunjin Kim, Minhyuk Sung
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 task adaptation (weight prediction plus SAM masks) to PartSLIP | PartSLIP · PartNet-Mobility semantic seg. | mIoU (%): 58.0 -> 65.0 | Improved |
| add bounding box weight prediction network (ablation) | PartSTAD · PartNet-Mobility semantic seg. | mIoU (%): 61.9 -> 65.0 | Improved |
| add SAM foreground mask integration (ablation) | PartSTAD · PartNet-Mobility semantic seg. | mIoU (%): 62.1 -> 65.0 | Improved |
| add weight prediction (ablation) | PartSTAD · PartNet-Mobility instance seg. | part-aware mAP50 (%): 44.7 -> 45.6 | Improved |
| use mRIoU loss instead of cross-entropy (ablation) | PartSTAD · PartNet-Mobility semantic seg. | mIoU (%): 63.5 -> 65.0 | Improved |
| combine cross-entropy with mRIoU loss (ablation) | PartSTAD · PartNet-Mobility semantic seg. | mIoU (%): 65.0 -> 64.5 | Hurt |
| render 10 views instead of 5 (ablation) | PartSTAD · PartNet-Mobility semantic seg. | mIoU (%): 59.4 -> 65.0 | Improved |
| increase training shapes from 8 to 128 (ablation) | PartSTAD · StorageFurniture | mIoU (%): 57.0 -> 60.1 | No clear effect |
| vary initial weight tau (ablation) | PartSTAD · PartNet-Mobility top-5 categories | mIoU (%): 54.8 (tau=1) vs 55.8 (tau=10) | No clear effect |
| use random rather than fixed viewpoints (ablation) | PartSTAD · Chair, Kettle, Lamp, Suitcase, Scissors | mIoU (%): Chair 85.3 vs 85.5; Suitcase 68.3 vs 71.3 | No clear effect |
| weight prediction applied to vanilla GLIP (ablation) | PartSTAD · PartNet-Mobility semantic seg. | mIoU (%): 27.2 -> 48.9 | Improved |
Closest papers by meaning
- PartDistill: 3D Shape Part Segmentation by Vision-Language Model Distillation (CVPR 2024 · similarity 0.87)
- SegGraph: Leveraging Graphs of SAM Segments for Few-Shot 3D Part Segmentation (NeurIPS 2025 · similarity 0.86)
- PartNeXt: A Next-Generation Dataset for Fine-Grained and Hierarchical 3D Part Understanding (NeurIPS 2025 · similarity 0.85)
- MLLM-For3D: Adapting Multimodal Large Language Model for 3D Reasoning Segmentation (NeurIPS 2025 · similarity 0.85)
- SAM2Object: Consolidating View Consistency via SAM2 for Zero-Shot 3D Instance Segmentation (CVPR 2025 · similarity 0.85)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (CVPR 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.
