Research map / Detection and segmentation / 3D detection and driving
Monocular 3D Object Detection with Decoupled Structured Polygon Estimation and Height-Guided Depth Estimation
Yingjie Cai, Buyu Li, Zeyu Jiao, Hongsheng Li, Xingyu Zeng, Xiaogang Wang
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 |
|---|---|---|---|
| decompose detection into structured polygon prediction and depth recovery rather than jointly regress all box variables (ablation) | Decoupled-3D · KITTI validation, moderate, IoU=0.5 | AP_BEV: 12.05 -> 54.31 | Improved |
| decompose detection into structured polygon prediction and depth recovery rather than jointly regress all box variables (ablation) | Decoupled-3D · KITTI validation, moderate, IoU=0.5 | AP_3D: 4.50 -> 50.50 | Improved |
| use structured polygon instead of BEV features to estimate coarse boxes (ablation) | Decoupled-3D · KITTI validation, moderate, IoU=0.5 | AP_BEV: 38.61 -> 54.31 | Improved |
| use structured polygon instead of BEV features to estimate coarse boxes (ablation) | Decoupled-3D · KITTI validation, moderate, IoU=0.5 | AP_3D: 27.88 -> 50.50 | Improved |
| use object-height-guided depth recovery instead of direct depth regression (ablation) | Decoupled-3D · KITTI validation | mean depth error: 2.41m -> 1.21m | Improved |
| add fine-grained refinement using bird's-eye-view spatial features (ablation) | Decoupled-3D coarse boxes · KITTI validation, moderate, IoU=0.5 | AP_3D: 16.36 -> 50.50 | Improved |
| refine coarse boxes with image-level features (ablation) | Decoupled-3D coarse boxes · KITTI validation, moderate, IoU=0.5 | AP_3D: 16.36 -> 16.95 | Improved |
| refine coarse boxes with front-view depth map (ablation) | Decoupled-3D coarse boxes · KITTI validation, moderate, IoU=0.5 | AP_3D: 16.36 -> 16.93 | Improved |
| refine coarse boxes with image-level features (ablation) | Decoupled-3D coarse boxes · KITTI validation, moderate, IoU=0.5 | AP_BEV: 20.91 -> 20.74 | No clear effect |
| refine coarse boxes with front-view depth map (ablation) | Decoupled-3D coarse boxes · KITTI validation, moderate, IoU=0.5 | AP_BEV: 20.91 -> 20.70 | No clear effect |
| add fine-grained refinement to Mono3D (ablation) | Mono3D · KITTI validation, moderate, IoU=0.5 | AP_3D: 18.20 -> 29.05 | Improved |
| add fine-grained refinement to MGR (ablation) | MGR · KITTI validation, moderate, IoU=0.5 | AP_3D: 36.97 -> 43.80 | Improved |
Closest papers by meaning
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.90)
- Objects Are Different: Flexible Monocular 3D Object Detection (CVPR 2021 · similarity 0.90)
- Geometry-Based Distance Decomposition for Monocular 3D Object Detection (ICCV 2021 · similarity 0.90)
- Rethinking Geometric Depth in Monocular 3D Object Detection: A Projection-Consistent Reformulation (NeurIPS 2026 · similarity 0.89)
- DID-M3D: Decoupling Instance Depth for Monocular 3D Object Detection (ECCV 2022 · similarity 0.89)
- MonoCD: Monocular 3D Object Detection with Complementary Depths (CVPR 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.
