Research map / Label-efficient and robust learning / Semi-supervised learning
All Points Matter: Entropy-Regularized Distribution Alignment for Weakly-supervised 3D Segmentation
Liyao Tang, Zhe Chen, Shanshan Zhao, Chaoyue Wang, Dacheng Tao
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 ERDA to RandLA-Net | RandLA-Net · S3DIS Area 5, 1% | mIoU: 59.8 -> 67.2 | Improved |
| Add ERDA to PT | PT · S3DIS Area 5, 1% | mIoU: 65.8 -> 70.4 | Improved |
| Add ERDA to CloserLook3D | CloserLook3D · ScanNet test, 20pts | mIoU: 54.4 -> 57.0 | Improved |
| Add entropy regularization to pseudo-label learning (ablation) | RandLA-Net with pseudo-labels · S3DIS Area 5, 1% | mIoU: 63.3 -> 65.1 | Improved |
| Add distribution alignment to pseudo-label learning (ablation) | RandLA-Net with pseudo-labels · S3DIS Area 5, 1% | mIoU: 63.3 -> 66.1 | Improved |
| Combine entropy regularization and distribution alignment (ablation) | RandLA-Net with pseudo-labels · S3DIS Area 5, 1% | mIoU: 63.3 -> 67.2 | Improved |
| Use KL(p || q) with λ=1 rather than KL(q || p) with λ=1 (ablation) | ERDA · S3DIS Area 5, 1% | mIoU: 65.9 -> 67.2 | Improved |
| Use KL(p || q) with λ=1 rather than JS with λ=1 (ablation) | ERDA · S3DIS Area 5, 1% | mIoU: 65.4 -> 67.2 | Improved |
| Use KL(p || q) with λ=1 rather than MSE with λ=1 (ablation) | ERDA · S3DIS Area 5, 1% | mIoU: 66.2 -> 67.2 | Improved |
| Increase pseudo-label coverage using ERDA rather than hard one-hot pseudo-labels (ablation) | RandLA-Net · S3DIS Area 5, 1% | mIoU: one-hot: 62.7; ERDA: 67.2 | Improved |
| Use soft pseudo-labels without ERDA instead of one-hot labels (ablation) | RandLA-Net · S3DIS Area 5, 1% | mIoU: dense one-hot: 62.7; dense soft: 62.6 | Hurt |
| Add a 2-layer MLP projection network (ablation) | ERDA · S3DIS Area 5, 1% | mIoU: 65.90 -> 67.18 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FixMatch — FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (NeurIPS 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
- Don’t Starve the Boundaries: Boundary-Constrained Label Propagation for Weakly Supervised 3D Segmentation (ECCV 2026 · similarity 0.90)
- W2P: Switching from Weak Supervision to Partial Supervision for Semantic Segmentation (AAAI 2024 · similarity 0.86)
- Fully Data-Driven Pseudo Label Estimation for Pointly-Supervised Panoptic Segmentation (AAAI 2024 · similarity 0.86)
- Towards Robust Pseudo-Label Learning in Semantic Segmentation: An Encoding Perspective (NeurIPS 2025 · similarity 0.86)
- Pseudo Label Refinery for Unsupervised Domain Adaptation on Cross-dataset 3D Object Detection (CVPR 2024 · similarity 0.86)
- Label-Efficient Dataset Pruning via Semi-Supervised Pseudo-Labeling (NeurIPS 2026 · 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.
