Research map / Detection and segmentation / 3D detection and driving
Diffusion Model for Robust Multi-Sensor Fusion in 3D Object Detection and BEV Segmentation
Duy Tho Le, Hengcan Shi, Jianfei Cai, Hamid Rezatofighi
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 |
|---|---|---|---|
| add cMini-BiFPN fusion architecture (ablation) | BEVFusion baseline · nuScenes val BEV segmentation | mIoU: 62.7 -> 66.5 | Improved |
| add cMini-BiFPN fusion architecture (ablation) | BEVFusion baseline · nuScenes val 3D detection | NDS: 71.4 -> 71.8 | Improved |
| add GSM module with PSDT (4-step sampling) (ablation) | cMini-BiFPN · nuScenes val BEV segmentation | mIoU: 66.5 -> 68.82 | Improved |
| increase diffusion sampling from 4 to 8 steps (ablation) | DifFUSER · nuScenes val BEV segmentation | mIoU: 68.82 -> 70.04 | Improved |
| add Shift to Scale in GSM (ablation) | GSM block · nuScenes val BEV segmentation (6-epoch) | mIoU: 59.92 -> 60.73 | Improved |
| add Gate to Scale and Shift in GSM (ablation) | GSM block · nuScenes val BEV segmentation (6-epoch) | mIoU: 60.73 -> 62.17 | Improved |
| swap DDIM for DPM-Solver++ or DEIS (ablation) | DifFUSER sampler · nuScenes val BEV segmentation | mIoU: 70.04 -> 70.02 / 69.95 | No clear effect |
| joint multi-task training instead of single-task (ablation) | DifFUSER · nuScenes val | NDS / mIoU: NDS 69.59 -> 68.23; mIoU 65.8 -> 61.36 | Hurt |
| replace BEVFusion with DifFUSER for fusion | BEVFusion · nuScenes val BEV segmentation | mIoU: 62.7 -> 70.04 | Improved |
| replace BEVFusion with DifFUSER for fusion | BEVFusion · nuScenes test 3D detection | NDS: 72.9 -> 73.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- FocalFormer3D — FocalFormer3D: Focusing on Hard Instance for 3D Object Detection (ICCV 2023) · better on 0 of 1 shared result
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
- BEVDiffuser: Plug-and-Play Diffusion Model for BEV Denoising with Ground-Truth Guidance (CVPR 2025 · similarity 0.88)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · similarity 0.86)
- DiffuBox: Refining 3D Object Detection with Point Diffusion (NeurIPS 2024 · similarity 0.86)
- DiMSOD: A Diffusion-Based Framework for Multi-Modal Salient Object Detection (AAAI 2025 · similarity 0.86)
- DiffBEV: Conditional Diffusion Model for Bird’s Eye View Perception (AAAI 2024 · similarity 0.86)
- DINTR: Tracking via Diffusion-based Interpolation (NeurIPS 2024 · similarity 0.85)
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.
