Research map / Point clouds / Segmentation and detection
LiDAR R-CNN: An Efficient and Universal 3D Object Detector
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 |
|---|---|---|---|
| Add LiDAR R-CNN refinement to PointPillars(TV) | PointPillars(TV) · KITTI testing set, vehicle Easy | 3D AP: 83.12 -> 85.97 | Helped |
| Add LiDAR R-CNN refinement to PointPillars* | PointPillars* · Waymo Open Dataset validation, vehicle LEVEL_1 | 3D AP: 72.1 -> 75.6 | Helped |
| Double the point-encoding channels (ablation) | LiDAR R-CNN · Waymo Open Dataset validation, vehicle LEVEL_1 | 3D AP: 75.6 -> 75.6 | No clear effect |
| Cascade another refinement stage (ablation) | LiDAR R-CNN (2x) · Waymo Open Dataset validation, vehicle LEVEL_1 | 3D AP: 75.6 -> 76.0 | Helped |
| Add boundary-offset point features (ablation) | PointNet refinement on PointPillars · Waymo Open Dataset validation, vehicle LEVEL_1 | 3D AP@70: 74.1 -> 75.6 | Helped |
| Add virtual points (ablation) | PointNet refinement on PointPillars · Waymo Open Dataset validation, vehicle LEVEL_1 | 3D AP@70: 74.1 -> 75.4 | Helped |
| Replace point-based refinement with voxelization (ablation) | PointNet refinement on PointPillars · Waymo Open Dataset validation, vehicle LEVEL_1 | 3D AP@70: 74.1 -> 72.9 | Hurt |
| Use size normalization (ablation) | PointNet refinement on PointPillars · Waymo Open Dataset validation, three classes | cyclist 3D AP: 64.4 -> 64.4 | No clear effect |
| Add boundary-offset point features (ablation) | PointNet refinement on PointPillars · Waymo Open Dataset validation, three classes | cyclist 3D AP: 64.4 -> 66.8 | Helped |
| Replace the RV voxel second stage with LiDAR R-CNN (ablation) | RV · Waymo Open Dataset validation, vehicle LEVEL_1 | 3D AP: 64.1 -> 68.7 | Helped |
| Add size regression using vanilla PointNet refinement (ablation) | PointPillars · Waymo Open Dataset validation, vehicle | AP@L1: 72.1 -> 72.2 | No clear effect |
| Add heading regression using vanilla PointNet refinement (ablation) | PointPillars · Waymo Open Dataset validation, vehicle | AP@L1: 72.1 -> 72.1 | No clear effect |
Closest papers by meaning
- Pseudo-LiDAR From Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving (CVPR 2019 · similarity 0.91)
- VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection (CVPR 2018 · similarity 0.91)
- LaserNet: An Efficient Probabilistic 3D Object Detector for Autonomous Driving (CVPR 2019 · similarity 0.91)
- Multimodal Virtual Point 3D Detection (NeurIPS 2021 · similarity 0.90)
- Fully Convolutional One-Stage 3D Object Detection on LiDAR Range Images (NeurIPS 2022 · similarity 0.90)
- An Empirical Study of the Generalization Ability of Lidar 3D Object Detectors to Unseen Domains (CVPR 2024 · similarity 0.89)
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.