Research map / Point clouds / Segmentation and detection
Localization and Expansion: A Decoupled Framework for Point Cloud Few-shot Semantic Segmentation
Zhaoyang Li, Yuan Wang, Wangkai Li, Rui Sun, Tianzhu Zhang
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 structural localization module (SLM) with agent-level correlation (ablation) | AttMPTI baseline · S3DIS S0 1-way 1-shot | mIoU: 66.27 -> 74.26 | Improved |
| add full self-expansion module (self-loss and cycle check) (ablation) | AttMPTI baseline · S3DIS S0 1-way 1-shot | mIoU: 66.27 -> 71.56 | Improved |
| add SEM on top of SLM (ablation) | DLE · S3DIS S0 1-way 1-shot | mIoU: 74.26 -> 76.54 | Improved |
| add cycle-consistency (conservative expansion) check (ablation) | AttMPTI baseline · S3DIS S0 1-way 1-shot | mIoU: 66.27 -> 70.61 | Improved |
| add self-loss only (ablation) | AttMPTI baseline · S3DIS S0 1-way 1-shot | mIoU: 66.27 -> 67.9 | Improved |
| initialize agents by clustering foreground support features instead of random selection (ablation) | SLM · S3DIS S0 1-way 1-shot | mIoU: 72.1 -> 76.54 | Improved |
| cluster foreground only versus all support points for agents (ablation) | SLM · S3DIS S0 1-way 1-shot | mIoU: 75.56 -> 76.54 | Improved |
| increase number of agents beyond 100 (ablation) | SLM · S3DIS S0 1-way 1-shot | mIoU: 76.54 -> 75.06 (200 agents) | Hurt |
| use DLE versus QGE on ScanNet 2-way 1-shot | DLE · ScanNet 2-way 1-shot | mIoU: +3.02% over best prior | Improved |
| DLE versus AttMPTI+QGE on S3DIS 2-way 5-shot S1 split | DLE · S3DIS 2-way 5-shot S1 | mIoU: 79.46 (QGE) vs 74.79 (Ours) | Hurt |
Closest papers by meaning
- Rethinking Few-shot 3D Point Cloud Semantic Segmentation (CVPR 2024 · similarity 0.87)
- Taylor Series-Inspired Local Structure Fitting Network for Few-shot Point Cloud Semantic Segmentation (AAAI 2025 · similarity 0.87)
- Generated and Pseudo Content guided Prototype Refinement for Few-shot Point Cloud Segmentation (NeurIPS 2024 · similarity 0.87)
- DyPolySeg: Taylor Series-Inspired Dynamic Polynomial Fitting Network for Few-shot Point Cloud Semantic Segmentation (ICML 2025 · similarity 0.86)
- Construct to Associate: Cooperative Context Learning for Domain Adaptive Point Cloud Segmentation (CVPR 2024 · similarity 0.85)
- Weakly Supervised Point Cloud Semantic Segmentation via Artificial Oracle (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.
