Research map / Point clouds / Segmentation and detection
Panoptic-PolarNet: Proposal-Free LiDAR Point Cloud Panoptic Segmentation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| share the first three decoding layers between semantic and instance heads (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 51.6% -> 52.3% | Helped |
| add instance oversampling (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: +2.8% | Helped |
| add instance global augmentation (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 55.1% -> 57.0% | Helped |
| use center heatmap sigma 5 (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 57.0% -> 57.3% | Helped |
| add instance local augmentation (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 57.0% -> 57.3% | Helped |
| combine instance oversampling, global augmentation, and local augmentation (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 57.3% -> 57.4% | Helped |
| add self-adversarial pruning (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 57.4% -> 57.5% | Helped |
| add visibility feature (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: +1.6% | Helped |
| use Cartesian rather than polar coordinates (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 59.1% -> 54.3% | Hurt |
| reduce the BEV grid size to [320,240,32] (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 59.1% -> 57.1% | Hurt |
| use focal loss rather than MSE for heatmap regression (ablation) | Panoptic-PolarNet · SemanticKITTI | PQ: -1.3% | Hurt |
| replace predicted heatmap and offsets with ground truth (ablation) | Panoptic-PolarNet · SemanticKITTI validation | PQ: 59.1% -> 60.1% | Helped |
Closest papers by meaning
- Panoptic-PHNet: Towards Real-Time and High-Precision LiDAR Panoptic Segmentation via Clustering Pseudo Heatmap (CVPR 2022 · similarity 0.94)
- PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation (CVPR 2020 · similarity 0.93)
- LiDAR-Based Panoptic Segmentation via Dynamic Shifting Network (CVPR 2021 · similarity 0.93)
- Beyond the Label Itself: Latent Labels Enhance Semi-supervised Point Cloud Panoptic Segmentation (AAAI 2024 · similarity 0.89)
- PC-BEV: An Efficient Polar-Cartesian BEV Fusion Framework for LiDAR Semantic Segmentation (AAAI 2025 · similarity 0.89)
- How Do Images Align and Complement LiDAR? Towards a Harmonized Multi-modal 3D Panoptic Segmentation (ICML 2025 · 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.