Research map / 3D vision / Depth and stereo
BEVStereo: Enhancing Depth Estimation in Multi-View 3D Object Detection with Temporal Stereo
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 dynamic temporal stereo to BEVDepth | BEVDepth · nuScenes test | mAP: 0.503 -> 0.525 | Helped |
| Add dynamic temporal stereo to BEVDepth | BEVDepth · nuScenes test | NDS: 0.600 -> 0.610 | Helped |
| Use sparse dynamic temporal stereo rather than dense MVS | BEVDepth · nuScenes val | memory: 24.04GB -> 8.01GB | Helped |
| Add dynamic temporal stereo to BEVDepth | BEVDepth · nuScenes val, moving objects (velocity > 1m/s), TH=0.5 | recall: 28.32 -> 29.79 | Helped |
| Add dynamic temporal stereo to BEVDepth | BEVDepth · nuScenes val, static ego vehicle (speed < 1 m/s) | mAP: 32.73 -> 33.76 | Helped |
| Add Weight Net to combine mono and stereo depth (ablation) | BEVStereo · nuScenes val | NDS: 44.7 -> 45.3 | Helped |
| Add Weight Net to combine mono and stereo depth (ablation) | BEVStereo · nuScenes val | mAP: 34.5 -> 34.6 | Helped |
| Increase μ iteration count from 0 to 3 (ablation) | BEVStereo · nuScenes val | mAP: 32.7 -> 34.6 | Helped |
| Increase μ iteration count from 0 to 3 (ablation) | BEVStereo · nuScenes val | NDS: 43.9 -> 45.3 | Helped |
| Use size-aware instead of conventional Circle NMS (ablation) | BEVStereo with class-aware NMS · nuScenes val, class-aware NMS | mAP: 34.6 -> 35.1 | Helped |
| Use size-aware instead of conventional Circle NMS (ablation) | BEVStereo with class-agnostic NMS · nuScenes val, class-agnostic NMS | mAP: 24.9 -> 33.3 | Helped |
| Use size-aware instead of conventional Circle NMS (ablation) | BEVStereo with class-agnostic NMS · nuScenes val, class-agnostic NMS | NDS: 38.0 -> 45.0 | Helped |
Closest papers by meaning
- Instance-Aware Multi-Camera 3D Object Detection with Structural Priors Mining and Self-Boosting Learning (AAAI 2024 · similarity 0.87)
- BEVNeXt: Reviving Dense BEV Frameworks for 3D Object Detection (CVPR 2024 · similarity 0.87)
- RaCFormer: Towards High-Quality 3D Object Detection via Query-based Radar-Camera Fusion (CVPR 2025 · similarity 0.86)
- GeoBEV: Learning Geometric BEV Representation for Multi-view 3D Object Detection (AAAI 2025 · similarity 0.85)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 2025 · similarity 0.85)
- PPMStereo: Pick-and-Play Memory Construction for Consistent Dynamic Stereo Matching (NeurIPS 2025 · 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.