Research map / Point clouds / Segmentation and detection
Pillar-based Object Detection for Autonomous Driving
Yue Wang, Alireza Fathi, Abhijit Kundu, David A. Ross, Caroline Pantofaru, Tom Funkhouser, Justin Solomon
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 |
|---|---|---|---|
| predict boxes per pillar instead of per anchor (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7), prediction head ablation | 3D mAP overall: 59.78 -> 69.8 | Improved |
| predict boxes per pillar instead of per anchor (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7), prediction head ablation | BEV mAP overall: 78.84 -> 87.11 | Improved |
| predict boxes per pillar instead of per point (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7), prediction head ablation | 3D mAP overall: 60.6 -> 69.8 | Improved |
| predict per point instead of per anchor (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7), prediction head ablation | 3D mAP overall: 59.78 -> 60.6 | Improved |
| add cylindrical view to BEV (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | 3D mAP overall: 61.86 -> 69.8 | Improved |
| use cylindrical view instead of spherical view in combination with BEV (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | 3D mAP overall: 66.31 -> 69.8 | Improved |
| use cylindrical view instead of XZ view in combination with BEV (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | 3D mAP overall: 62.76 -> 69.8 | Improved |
| add spherical view to BEV (as in MVF) (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | 3D mAP overall: 61.86 -> 66.31 | Improved |
| add XZ view to BEV (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | 3D mAP overall: 61.86 -> 62.76 | No clear effect |
| single cylindrical view instead of single BEV (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | 3D mAP 50m-Inf: 33.53 -> 37.99 | Improved |
| use bilinear instead of nearest neighbor interpolation in pillar-to-point projection (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | 3D mAP overall: 64.76 -> 69.8 | Improved |
| use bilinear instead of nearest neighbor interpolation in pillar-to-point projection (ablation) | Pillar-based detector · Waymo vehicle (IoU=0.7) | BEV mAP 50m-Inf: 65.77 -> 72.12 | Improved |
Closest papers by meaning
- PillarNet: Real-Time and High-Performance Pillar-Based 3D Object Detection (ECCV 2022 · similarity 0.92)
- PillarHist: A Quantization-aware Pillar Feature Encoder based on Height-aware Histogram (CVPR 2025 · similarity 0.89)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 2024 · similarity 0.87)
- OpenAD: Open-World Autonomous Driving Benchmark for 3D Object Detection (NeurIPS 2025 · similarity 0.86)
- SwiftPillars: High-Efficiency Pillar Encoder for Lidar-Based 3D Detection (AAAI 2024 · similarity 0.86)
- CorrBEV: Multi-View 3D Object Detection by Correlation Learning with Multi-modal Prototypes (CVPR 2025 · similarity 0.86)
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.
