Research map / Point clouds / Segmentation and detection
Multi-View 3D Object Detection Network for Autonomous Driving
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 deep fusion instead of early or late fusion (ablation) | MV3D region-based fusion network · KITTI validation car moderate | AP_3D at IoU=0.5: 87.60 -> 88.29; 87.70 -> 88.29 | Helped |
| add auxiliary losses to deep fusion (ablation) | MV3D region-based fusion network · KITTI validation car moderate | AP_3D at IoU=0.5: 88.29 -> 89.05 | Helped |
| add RGB features to BV+FV (ablation) | MV3D · KITTI validation car moderate | AP_3D at IoU=0.5: 87.65 -> 89.05 | Helped |
| combine any two views rather than using one view alone (ablation) | MV3D multi-view feature network · KITTI validation car moderate | AP_3D at IoU=0.5 | Helped |
| use bird's-eye-view rather than front-view features alone (ablation) | MV3D · KITTI validation car moderate | AP_3D at IoU=0.5: 56.30 -> 85.50 | Helped |
| use BV+FV rather than VeloFCN | KITTI 3D detection · KITTI validation car moderate | AP_3D at IoU=0.5: 57.57 -> 87.65; ~30% | Helped |
| add RGB images to the LIDAR-based BV+FV model | MV3D · KITTI validation car moderate | AP_loc at IoU=0.5: 88.57 -> 89.39 | Helped |
| use BV+FV rather than VeloFCN | KITTI 3D localization · KITTI validation car moderate, IoU=0.5 | AP_loc: 63.82 -> 88.57; ~25% | Helped |
| generate proposals from the LIDAR bird's-eye-view representation | MV3D proposal network · KITTI validation car moderate, 300 proposals | 3D recall at IoU=0.5: 91%; 73.9% | Helped |
| optimize 3D boxes and project them for 2D detection | MV3D (BV+FV) · KITTI test car hard | AP_2D: 68.30 -> 78.61; 10.31% | Helped |
Closest papers by meaning
- LaserNet: An Efficient Probabilistic 3D Object Detector for Autonomous Driving (CVPR 2019 · similarity 0.92)
- A Versatile Multi-View Framework for LiDAR-Based 3D Object Detection With Guidance From Panoptic Segmentation (CVPR 2022 · similarity 0.91)
- Multimodal Virtual Point 3D Detection (NeurIPS 2021 · similarity 0.91)
- Pseudo-LiDAR From Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving (CVPR 2019 · similarity 0.91)
- SAFDNet: A Simple and Effective Network for Fully Sparse 3D Object Detection (CVPR 2024 · similarity 0.89)
- BEV-MAE: Bird’s Eye View Masked Autoencoders for Point Cloud Pre-training in Autonomous Driving Scenarios (AAAI 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.