Research map / Point clouds / Segmentation and detection
GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloud
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 |
|---|---|---|---|
| Use R-PointNet instead of SGPN-CNN | NYUv2 instance segmentation · NYUv2 | AP: 33.6 -> 39.3 | Helped |
| Use R-PointNet instead of SGPN | PartNet part instance segmentation · PartNet Chair | AP: 0.194 -> 0.268 | Helped |
| Replace direct box regression proposals with GSPN (ablation) | R-PointNet · ScanNet v2 validation | AP: 15.8 -> 19.3 | Helped |
| Replace binary-segmentation-based proposals with GSPN (ablation) | R-PointNet · ScanNet v2 validation | AP: 14.9 -> 19.3 | Helped |
| Use CVAE rather than a naive encoder-decoder (ablation) | GSPN · ScanNet v2 validation | mIoU: 0.408 -> 0.581 | Helped |
| Use multi-scale rather than single-scale context (ablation) | GSPN · ScanNet v2 validation | mIoU: 0.486 -> 0.581 | Helped |
| Predict the object center before generation (ablation) | GSPN · ScanNet v2 validation | mIoU: 0.409 -> 0.581 | Helped |
| Include semantic feature f_sem (ablation) | R-PointNet feature backbone · ScanNet v2 validation | AP: 0.161 -> 0.191 | Helped |
| Include instance-sensitive context feature f_hat_c (ablation) | R-PointNet feature backbone · ScanNet v2 validation | AP: 0.178 -> 0.191 | Helped |
| Pretrain the semantic feature network (ablation) | R-PointNet feature backbone · ScanNet v2 validation | AP: 0.180 -> 0.191 | Helped |
| Use GSPN rather than direct box regression proposals (ablation) | R-PointNet · ScanNet v2 validation | mIoU: 0.514 -> 0.581 | Helped |
Closest papers by meaning
- FreePoint: Unsupervised Point Cloud Instance Segmentation (CVPR 2024 · similarity 0.87)
- Relation3D : Enhancing Relation Modeling for Point Cloud Instance Segmentation (CVPR 2025 · similarity 0.87)
- Point-SAM: Promptable 3D Segmentation Model for Point Clouds (ICLR 2025 · similarity 0.86)
- OneFormer3D: One Transformer for Unified Point Cloud Segmentation (CVPR 2024 · similarity 0.86)
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.86)
- Point Cloud Part Editing: Segmentation, Generation, Assembly, and Selection (AAAI 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.