Research map / Detection and segmentation / 3D detection and driving
Virtual Sparse Convolution for Multimodal 3D Object Detection
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 VirConvNet to Voxel-RCNN (ablation) | Virtual-point-only detection on KITTI validation · KITTI validation, virtual points only | 3D AP: 3.43% | Helped |
| Add VirConvNet to early fusion (ablation) | Voxel-RCNN early-fusion detection · KITTI validation, early fusion | 3D AP: 2.93% | Helped |
| Add VirConvNet to late fusion (ablation) | Voxel-RCNN late-fusion detection · KITTI validation, late fusion | 3D AP: 2.65% | Helped |
| Add StVD (ablation) | VirConv-T · KITTI validation, car | 3D AP, moderate: 87.55 -> 88.32 | Helped |
| Add NRConv to StVD (ablation) | VirConv-T · KITTI validation, car | 3D AP, moderate: 88.32% to 90.29% | Helped |
| Add TRS (ablation) | Virtual-point-only Voxel-RCNN with VirConv · KITTI validation, virtual points only | 3D AP: 1.36% | Helped |
| Add TRS (ablation) | Early-fusion Voxel-RCNN with VirConv · KITTI validation, early fusion | 3D AP: 0.25% | Helped |
| Add TRS (ablation) | Late-fusion Voxel-RCNN with VirConv · KITTI validation, late fusion | 3D AP: 1.32% | Helped |
| Add VirConv to CenterPoint with virtual points | CenterPoint + VP · nuScenes test | mAP: 66.4 -> 67.2 | Helped |
| Add VirConv to TransFusion-L with virtual points | TransFusion-L + VP · nuScenes test | NDS: 70.8 -> 72.3 | Helped |
| Train VirConv-T using labeled and pseudo-labeled data | VirConv-S · KITTI validation, car | 3D AP (R40), moderate: 90.29 -> 90.97 | Helped |
Closest papers by meaning
- ViKIENet: Towards Efficient 3D Object Detection with Virtual Key Instance Enhanced Network (CVPR 2025 · similarity 0.91)
- MsSVT: Mixed-scale Sparse Voxel Transformer for 3D Object Detection on Point Clouds (NeurIPS 2022 · similarity 0.89)
- HVPR: Hybrid Voxel-Point Representation for Single-Stage 3D Object Detection (CVPR 2021 · similarity 0.88)
- SAFDNet: A Simple and Effective Network for Fully Sparse 3D Object Detection (CVPR 2024 · similarity 0.86)
- Voxel or Pillar: Exploring Efficient Point Cloud Representation for 3D Object Detection (AAAI 2024 · similarity 0.86)
- FSHNet: Fully Sparse Hybrid Network for 3D Object Detection (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.