Research map / Point clouds / Segmentation and detection
InSeg: Interactive Refinement via Intent Propagation for Point Cloud Semantic Segmentation
Yudong Liu, Yunfei Li, Ge Gao, Han Huang, Ming Gu
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 the feature similarity mask to point-wise click supervision (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 75.10 -> 78.89 mIoU | Improved |
| add the spatial connectivity mask on top of the feature mask (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 78.89 -> 81.39 mIoU | Improved |
| add the instance prior mask on top of feature and connectivity masks (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 81.39 -> 83.62 mIoU | Improved |
| add the history replay loss (gamma 0 to 2) (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 82.20 -> 83.62 mIoU | Improved |
| increase replay loss weight from 2 to 4 (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 83.62 -> 82.91 | Hurt |
| exclude the current intent-consistent region from entropy minimization (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 82.91 -> 83.62 mIoU | Improved |
| add similarity virtual clicks (Nsim 0 to 15) (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 82.62 -> 83.62 | Improved |
| add region virtual clicks (Nreg 0 to 5) (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 82.95 -> 83.62 | Improved |
| increase similarity virtual clicks from 15 to 20 (ablation) | InSeg · ScanNet mIoU@5 clicks | mIoU: 83.62 -> 82.87 | Hurt |
| replace point-wise ISSR supervision with intent propagation (InSeg) | SpUNet on ScanNet · ScanNet mIoU@5 clicks | mIoU: 75.1 -> 83.6 | Improved |
| replace ISSR with InSeg | RandLA-Net on SemanticKITTI · SemanticKITTI mIoU@5 clicks | mIoU: 53.1 -> 62.1 | Improved |
| replace ISSR with InSeg | SpUNet on ScanNet · ScanNet completion rate NoC@75 within 5 clicks | completion rate (%): 57.5% to 78.5% | Improved |
Closest papers by meaning
- Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusion (CVPR 2021 · similarity 0.87)
- EPSegFZ: Efficient Point Cloud Semantic Segmentation for Few- and Zero-Shot Scenarios with Language Guidance (AAAI 2026 · similarity 0.87)
- PARI: Policy-Driven Active Residual Intervention for Weakly Supervised Point Cloud Segmentation (NeurIPS 2026 · similarity 0.87)
- Diffusion-Based Contextual Reconstruction for Point Cloud Segmentation with Limited Annotations (AAAI 2026 · similarity 0.86)
- From Coarse to Fine: Deep Prototype Refinement Network for Few-Shot Point Cloud Semantic Segmentation (ICML 2026 · similarity 0.86)
- Generated and Pseudo Content guided Prototype Refinement for Few-shot Point Cloud Segmentation (NeurIPS 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.
