Research map / Detection and segmentation / 3D detection and driving
PillarNet: Real-Time and High-Performance Pillar-Based 3D Object Detection
Guangsheng Shi, Ruifeng Li, Chao Ma
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 |
|---|---|---|---|
| replace PointPillars encoder with PillarNet-18 sparse 2D conv encoder | CenterPoint-PointPillars · nuScenes val | mAP: +7.61% mAP and +5.94% NDS | Improved |
| add dense-conv semantic stage-5 neck fusion (neckv2 over neckv1) (ablation) | PillarNet-18 · nuScenes val | NDS: 66.16 -> 66.41 | Improved |
| use neckv3 over neckv1 (ablation) | PillarNet-18 · nuScenes val | NDS: 66.16 -> 67.15 | Improved |
| use dense convolutions instead of sparse in neckv2 (neckv2-D) (ablation) | PillarNet-18 · nuScenes val | NDS: 66.41 -> 66.96 | Improved |
| add OD-DIoU regression loss (ablation) | PillarNet-18 · nuScenes val | mAP: +0.24% mAP | Improved |
| add IoU-aware rectification (ablation) | PillarNet-18 · nuScenes val | mAP: +0.34% | Improved |
| increase pillar size from 0.075m to 0.075*8m (ablation) | PillarNet-18 · nuScenes val | NDS: 67.15 -> 64.20 | Hurt |
| increase pillar size from 0.075m to 0.075*2m (ablation) | PillarNet-18 · nuScenes val | mAP: 59.48 -> 58.70 | Hurt |
| scale backbone from vgg to 34 (ablation) | PillarNet · nuScenes val | NDS: 65.71 -> 67.50 | Improved |
| use fine 0.075m pillars with lightweight PointPillars encoder (ablation) | PointPillars · nuScenes val | NDS: 60.22 -> 59.51 | Hurt |
| use two-frame input instead of single frame | PillarNet-18 · Waymo val Vehicle L2 | mAPH: 69.92 -> 71.08 | Improved |
| use two-frame input for cyclist | PillarNet-34 · Waymo val Cyclist L1 | mAP: 72.29 -> 70.51 | Hurt |
Later papers that compared with this method
1 later accepted paper reports PillarNet-34 as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- PillarHist: A Quantization-aware Pillar Feature Encoder based on Height-aware Histogram (CVPR 2025 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AFDetV2 — AFDetV2: Rethinking the Necessity of the Second Stage for Object Detection from Point Clouds (AAAI 2022) · better on 5 of 5 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
- Pillar-based Object Detection for Autonomous Driving (ECCV 2020 · similarity 0.92)
- PillarHist: A Quantization-aware Pillar Feature Encoder based on Height-aware Histogram (CVPR 2025 · similarity 0.90)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 2024 · similarity 0.89)
- SwiftPillars: High-Efficiency Pillar Encoder for Lidar-Based 3D Detection (AAAI 2024 · similarity 0.88)
- SAFDNet: A Simple and Effective Network for Fully Sparse 3D Object Detection (CVPR 2024 · similarity 0.86)
- DALDet: Depth-Aware Learning Based Object Detection for Autonomous Driving (AAAI 2024 · 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.
