Research map / Point clouds / Segmentation and detection
Hierarchical Aggregation for 3D Instance Segmentation
Shaoyu Chen, Jiemin Fang, Qian Zhang, Wenyu Liu, Xinggang Wang
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 intra-instance prediction network (ablation) | HAIS · ScanNet v2 val | AP50: 61.0 -> 63.4 | Improved |
| add intra-instance prediction network (ablation) | HAIS · ScanNet v2 val | AP: 39.8 -> 42.5 | Improved |
| add intra-instance prediction network (ablation) | HAIS · ScanNet v2 val | AP25: 74.6 -> 74.9 | Improved |
| add set aggregation (ablation) | HAIS · ScanNet v2 val | AP50: 63.4 -> 64.1 | Improved |
| add set aggregation (ablation) | HAIS · ScanNet v2 val | AP: 42.5 -> 43.5 | Improved |
| add set aggregation (ablation) | HAIS · ScanNet v2 val | AP25: 74.9 -> 75.6 | Improved |
| use masks to filter background points and compute IoU supervision for scoring (ablation) | HAIS · ScanNet v2 val | AP: 41.9 -> 43.5 | Improved |
| use masks to filter background points and compute IoU supervision for scoring (ablation) | HAIS · ScanNet v2 val | AP50: 63.1 -> 64.1 | Improved |
| filter mask training samples to IoU > 0.5 instead of using all (ablation) | HAIS · ScanNet v2 val | AP: 42.8 -> 43.5 | Improved |
| filter mask training samples to IoU > 0.5 instead of using all (ablation) | HAIS · ScanNet v2 val | AP25: 74.4 -> 75.6 | Improved |
| use HAIS instead of previous SOTA on test set | HAIS · ScanNet v2 test | AP50: 63.8 -> 69.9 | Improved |
| use HAIS instead of PointGroup | HAIS · ScanNet v2 val | per-frame inference time (ms): 452 -> 410 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- PointGroup — PointGroup: Dual-Set Point Grouping for 3D Instance Segmentation (CVPR 2020) · better on 7 of 7 shared results
- 3D-MPA — 3D-MPA: Multi-Proposal Aggregation for 3D Semantic Instance Segmentation (CVPR 2020) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Relation3D : Enhancing Relation Modeling for Point Cloud Instance Segmentation (CVPR 2025 · similarity 0.88)
- Insightful Instance Features for 3D Instance Segmentation (CVPR 2025 · similarity 0.88)
- FreePoint: Unsupervised Point Cloud Instance Segmentation (CVPR 2024 · similarity 0.88)
- Any3DIS: Class-Agnostic 3D Instance Segmentation by 2D Mask Tracking (CVPR 2025 · similarity 0.87)
- Spherical Mask: Coarse-to-Fine 3D Point Cloud Instance Segmentation with Spherical Representation (CVPR 2024 · similarity 0.87)
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.87)
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.
