Research map / Point clouds / Segmentation and detection
PI-RCNN: An Efficient Multi-Sensor 3D Object Detector with Point-Based Attentive Cont-Conv Fusion Module
Read the paperCheck 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 |
|---|---|---|---|
| Add PI-RCNN multi-sensor fusion to PointRCNN | PointRCNN · KITTI val Car | 3D AP Moderate: 77.10 -> 78.53 | Helped |
| Add PI-RCNN multi-sensor fusion to PointRCNN | PointRCNN · KITTI test | 3D AP Moderate: 74.59 -> 74.82 | Helped |
| Add PACF fusion (ablation) | PointRCNN · KITTI val Car | 3D AP Moderate: 77.10 -> 77.96 | Helped |
| Add PointPool to PACF, without Att Aggr (ablation) | PACF · KITTI val Car | 3D AP Moderate: 77.96 -> 78.42 | Helped |
| Add Att Aggr to PACF, without PointPool (ablation) | PACF · KITTI val Car | 3D AP Moderate: 77.96 -> 78.22 | Helped |
| Add both PointPool and Att Aggr (ablation) | PACF · KITTI val Car | 3D AP Moderate: 77.96 -> 78.53 | Helped |
| Set K to 3 instead of K=1 (ablation) | PI-RCNN · KITTI val Car | 3D AP Moderate: 77.59 -> 78.53 | Helped |
| Set K to 5 instead of K=3 (ablation) | PI-RCNN · KITTI val Car | 3D AP Moderate: 78.53 -> 77.98 | Hurt |
| Set K to 10 instead of K=3 (ablation) | PI-RCNN · KITTI val Car | 3D AP Moderate: 78.53 -> 77.15 | Hurt |
| Move fusion from the middle to the beginning of the detector (ablation) | PI-RCNN V1 · KITTI val Car | 3D AP Moderate: 78.53 -> 78.01 | Hurt |
| Train segmentation with all categories rather than only the target class (ablation) | PI-RCNN · KITTI val Car | 3D AP Moderate: 77.16 -> 78.53 | Helped |
| Pretrain the segmentation sub-network (ablation) | PI-RCNN · KITTI val Car | 3D AP Moderate: 76.34 -> 77.16 | Helped |
Closest papers by meaning
- PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection (CVPR 2020 · similarity 0.88)
- Multi-View 3D Object Detection Network for Autonomous Driving (CVPR 2017 · similarity 0.88)
- Multi-Task Multi-Sensor Fusion for 3D Object Detection (CVPR 2019 · similarity 0.88)
- LoGoNet: Towards Accurate 3D Object Detection With Local-to-Global Cross-Modal Fusion (CVPR 2023 · similarity 0.88)
- Uni3D: A Unified Baseline for Multi-Dataset 3D Object Detection (CVPR 2023 · similarity 0.88)
- ViKIENet: Towards Efficient 3D Object Detection with Virtual Key Instance Enhanced Network (CVPR 2025 · similarity 0.87)
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.