Research map / Detection and segmentation / 3D detection and driving
MSMDFusion: Fusing LiDAR and Camera at Multiple Scales With Multi-Depth Seeds for 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 multi-depth unprojection and GMA-Conv (ablation) | TransFusion-L · nuScenes validation, 1/4 training split | mAP: 60.26 -> 66.93 | Helped |
| Add depth-aware feature generation to MDU (ablation) | MSMDFusion · nuScenes validation, 1/4 training split | mAP: 64.33 -> 65.59 | Helped |
| Add GMA-Conv to select useful camera voxel information (ablation) | MSMDFusion with MDU · nuScenes validation, 1/4 training split | mAP: 65.59 -> 66.93 | Helped |
| Perform interaction at more voxel scales, from C2-C3 to C2-C5 (ablation) | MSMDFusion · nuScenes validation, 1/4 training split | mAP: 66.34 -> 66.93 | Helped |
| Increase the number of depths per seed from 1 to 6 (ablation) | MSMDFusion · nuScenes validation, 1/4 training split | mAP: 65.86 -> 66.93 | Helped |
| Increase the number of depths per seed from 6 to 10 (ablation) | MSMDFusion · nuScenes validation, 1/4 training split | mAP: 66.93 -> 66.61 | Hurt |
| Increase seeds per instance from 50 to 200 while using one depth (ablation) | MSMDFusion · nuScenes validation, 1/4 training split | mAP: 65.86 -> 65.59 | Hurt |
| Use the full multi-scale fusion framework | MSMDFusion · nuScenes test | mAP: 70.8 -> 71.5 | Helped |
Closest papers by meaning
- DeepFusion: Lidar-Camera Deep Fusion for Multi-Modal 3D Object Detection (CVPR 2022 · similarity 0.92)
- MSeg3D: Multi-Modal 3D Semantic Segmentation for Autonomous Driving (CVPR 2023 · similarity 0.90)
- MsSVT: Mixed-scale Sparse Voxel Transformer for 3D Object Detection on Point Clouds (NeurIPS 2022 · similarity 0.89)
- GAFusion: Adaptive Fusing LiDAR and Camera with Multiple Guidance for 3D Object Detection (CVPR 2024 · similarity 0.88)
- Multi-View 3D Object Detection Network for Autonomous Driving (CVPR 2017 · similarity 0.88)
- SimDistill: Simulated Multi-Modal Distillation for BEV 3D Object Detection (AAAI 2024 · similarity 0.88)
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.