Research map / Detection and segmentation / 3D detection and driving
CHARM3R: Towards Unseen Camera Height Robust Monocular 3D Detector
Abhinav Kumar, Yuliang Guo, Zhihao Zhang, Xinyu Huang, Liu Ren, Xiaoming Liu
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 |
|---|---|---|---|
| average ground and regressed depth within the model (CHARM3R) | GUP Net · CARLA Val dH=+0.76m | AP3D70: 7.23 -> 27.33 | Improved |
| average ground and regressed depth within the model (CHARM3R) | GUP Net · CARLA Val dH=-0.70m | AP3D70: 9.46 -> 19.45 | Improved |
| average ground and regressed depth within the model (CHARM3R) | DEVIANT · CARLA Val dH=+0.76m | AP3D70: 6.25 -> 26.24 | Improved |
| average ground and regressed depth within the model (CHARM3R) | GUP Net · CARLA Val dH=0m (in-domain) | AP3D50: 76.47 -> 74.47 | Hurt |
| add ground depth to regressed depth (vs regressed only) (ablation) | CHARM3R GUP Net · CARLA Val dH=+0.76m | AP3D70: 7.23 -> 27.33 | Improved |
| add regressed depth to ground depth (vs ground only) (ablation) | CHARM3R GUP Net · CARLA Val dH=0m | AP3D70: 26.61 -> 55.68 | Improved |
| merge depths within model instead of offline (ablation) | CHARM3R GUP Net · CARLA Val dH=+0.76m | AP3D70: 18.36 -> 27.33 | Improved |
| use simple average instead of learned average (ablation) | CHARM3R GUP Net · CARLA Val dH=+0.76m | AP3D70: 9.53 -> 27.33 | Improved |
| add ReLU on ground depth (ablation) | CHARM3R GUP Net · CARLA Val dH=+0.76m | AP3D70: 0.07 -> 27.33 | Improved |
| use product formulation of alpha instead of sum (ablation) | CHARM3R GUP Net · CARLA Val dH=0m | AP3D70: 37.22 -> 55.68 | Improved |
| learned average instead of simple average (ablation) | CHARM3R GUP Net · CARLA Val dH=0m (in-domain) | AP3D70: 55.68 -> 56.49 | Hurt |
| CHARM3R on ResNet-18 backbone | GUP Net · CARLA Val dH=+0.76m | AP3D70: 5.28 -> 24.50 | Improved |
Closest papers by meaning
- Rethinking Geometric Depth in Monocular 3D Object Detection: A Projection-Consistent Reformulation (NeurIPS 2026 · similarity 0.89)
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.88)
- MonoCD: Monocular 3D Object Detection with Complementary Depths (CVPR 2024 · similarity 0.87)
- RefineAny3D: Depth Refinement as Semantic Alignment for Monocular 3D Detection (NeurIPS 2026 · similarity 0.86)
- Pow3R: Empowering Unconstrained 3D Reconstruction with Camera and Scene Priors (CVPR 2025 · similarity 0.86)
- MonoPlace3D: Learning 3D-Aware Object Placement for 3D Monocular Detection (CVPR 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.
