Research map / Point clouds / Segmentation and detection
PolarStream: Streaming Object Detection and Segmentation with Polar Pillars
Qi Chen, Sourabh Vora, Oscar Beijbom
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 bidirectional context padding (ablation) | Polar Pillars streaming model · nuScenes val, 32 sectors | PQ: 60.3 -> 68.3 | Improved |
| Add bidirectional context padding (ablation) | Polar Pillars streaming model · nuScenes val, 32 sectors | Segmentation mIoU: 67.1 -> 73.1 | Improved |
| Add bidirectional context padding (ablation) | Polar Pillars streaming model · nuScenes val, 32 sectors | Detection mAP: 46.7 -> 52.4 | Improved |
| Add trailing-edge context padding (ablation) | Polar Pillars streaming model · nuScenes val, 4 sectors | PQ: 67.7 -> 69 | Improved |
| Add trailing-edge context padding (ablation) | Polar Pillars streaming model · nuScenes val, 2 sectors | Detection mAP: 52.1 -> 52.2 | No clear effect |
| Add range stratified convolution and normalization (ablation) | Polar Pillars detector · nuScenes val, full sweep | Detection mAP: 48.2 -> 49.1 | Improved |
| Add feature undistortion (ablation) | Polar Pillars detector · nuScenes val, full sweep | Detection mAP: 48.2 -> 48.6 | Improved |
| Add both range stratified convolution and normalization and feature undistortion (ablation) | Polar Pillars detector · nuScenes val, full sweep | Detection mAP: 48.2 -> 50.3 | Improved |
| Add range stratified convolution and normalization and feature undistortion (ablation) | Cartesian Pillars detector · nuScenes val, full sweep | Detection mAP: 50.6 -> 50.6 | No clear effect |
| Add joint semantic segmentation training (ablation) | Polar Pillars detector with both polar-grid techniques · nuScenes val, full sweep | Detection mAP: 50.3 -> 51.2 | Improved |
| Add joint semantic segmentation training (ablation) | Cartesian Pillars · nuScenes val, full sweep | Segmentation mIoU: 72.5 -> 72.1 | Hurt |
| Use streaming sectors rather than full-sweep input | Polar Pillars · nuScenes val, 2, 4, 8, and 16 sectors | Detection performance | Improved |
Later papers that compared with this method
2 later accepted papers report PolarStream as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- PARTNER: Level Up the Polar Representation for 3D Object Detection (IJCV 2026 · better on 4 of 4 shared results)
- PARTNER: Level up the Polar Representation for LiDAR 3D Object Detection (ICCV 2023 · better on 2 of 2 shared results)
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
- PARTNER: Level up the Polar Representation for LiDAR 3D Object Detection (ICCV 2023 · similarity 0.92)
- PARTNER: Level Up the Polar Representation for 3D Object Detection (IJCV 2026 · similarity 0.89)
- Streaming Object Detection for 3-D Point Clouds (ECCV 2020 · similarity 0.89)
- PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation (CVPR 2020 · similarity 0.88)
- PC-BEV: An Efficient Polar-Cartesian BEV Fusion Framework for LiDAR Semantic Segmentation (AAAI 2025 · similarity 0.86)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 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.
