Research map / Point clouds / Segmentation and detection
Learning Semantic Segmentation of Large-Scale Point Clouds with Random Sampling
Qingyong Hu, Bo Yang, Linhai Xie, Stefano Rosa, Yulan Guo, Zhihua Wang, Niki Trigoni, Andrew Markham
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 local spatial encoding (LocSE) (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 49.8 -> 57.1 | Improved |
| use attentive pooling instead of mean-pooling (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 53.4 -> 57.1 | Improved |
| use attentive pooling instead of max-pooling (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 55.2 -> 57.1 | Improved |
| chain two LocSE+attentive pooling units in dilated residual block instead of a single one (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 48.8 -> 57.1 | Improved |
| use two sets instead of one in dilated residual block (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 52.9 -> 57.1 | Improved |
| add a third LocSE+attentive pooling set to dilated residual block (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 57.1 -> 54.2 | Hurt |
| encode all spatial information (p_i, p_i^k, relative position, distance) instead of p_i only (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 48.9 -> 57.1 | Improved |
| add point position, neighbour position and distance to relative position in LocSE (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 56.4 -> 57.1 | Improved |
| use random sampling instead of inverse density importance sampling (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 46.8 -> 57.1 | Improved |
| use random sampling instead of farthest point sampling (ablation) | RandLA-Net · SemanticKITTI val and S3DIS Area-5 | mIoU (%): 56.5 -> 57.1 on SemanticKITTI; 64.3 -> 63.4 on S3DIS | Mixed |
| add RGB colour input | RandLA-Net · Toronto-3D | mIoU (%): 77.71 -> 81.77 | Improved |
| increase K in KNN from 4 to 16 (ablation) | RandLA-Net · SemanticKITTI val (seq 08) | mIoU (%): 46% -> 57% | 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.
- PointASNL — PointASNL: Robust Point Clouds Processing Using Nonlocal Neural Networks With Adaptive Sampling (CVPR 2020) · better on 1 of 1 shared result
- PolarNet — PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic 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
- RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds (CVPR 2020 · similarity 0.99)
- Deep FusionNet for Point Cloud Semantic Segmentation (ECCV 2020 · similarity 0.90)
- Point-MoE: Large-Scale Multi-Dataset Training with Mixture-of-Experts for 3D Semantic Segmentation (ICLR 2026 · similarity 0.90)
- DeepLA-Net: Very Deep Local Aggregation Networks for Point Cloud Analysis (CVPR 2025 · similarity 0.89)
- Weakly Supervised Point Cloud Semantic Segmentation via Artificial Oracle (CVPR 2024 · similarity 0.88)
- Geometrically-driven Aggregation for Zero-shot 3D Point Cloud Understanding (CVPR 2024 · similarity 0.88)
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.
