Research map / Detection and segmentation / 3D detection and driving
Unleashing the Power of Chain-of-Prediction for Monocular 3D Object Detection
Zhihao Zhang, Abhinav Kumar, Girish Chandar Ganesan, 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 |
|---|---|---|---|
| add Chain-of-Prediction (feature-level chain) to the baseline (ablation) | MonoCoP baseline · KITTI Val Car (IoU 0.7) | AP3D Moderate: 21.12 -> 23.64 | Improved |
| add Uncertainty-Guided Selector on top of CoP (ablation) | MonoCoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 23.64 -> 23.98 (Moderate); 32.40 -> 32.06 (Easy) | Mixed |
| use GS routing instead of random routing (ablation) | MonoCoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 22.35 -> 23.98 | Improved |
| add feature propagation to feature learning in CoP (ablation) | CoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 21.74 -> 22.22 | Improved |
| add residual feature aggregation to CoP (ablation) | CoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 22.22 -> 23.98 | Improved |
| predict size, then orientation, then depth rather than other orders (ablation) | CoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 22.54 -> 23.98 vs depth-first; 23.08 -> 23.98 vs orientation-first | Improved |
| model attribute correlations with CoP instead of HTL (ablation) | MonoCoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 18.42 -> 23.98 | Improved |
| use larger backbone ResNet-101 instead of ResNet-50 (ablation) | MonoCoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 23.98 -> 21.75 | Hurt |
| use smaller backbone ResNet-34 instead of ResNet-50 (ablation) | MonoCoP · KITTI Val Car (IoU 0.7) | AP3D Moderate: 23.98 -> 22.32 | Hurt |
| switch from MonoDGP to MonoCoP (adds modules) | MonoDGP · KITTI Val Car (IoU 0.7) | parameters (M): 38.90 -> 42.50 params; 22.34 -> 23.98 AP3D | Mixed |
| MonoCoP vs best prior on Waymo 0.7 IoU Level 1 All APH3D | MonoCoP · Waymo Val Vehicle Level 1 (IoU 0.7) | APH3D All: 3.16 (MonoUNI) vs 2.70 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 5 on most shared results.
- MonoDGP — MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025) · better on 4 of 5 shared results
- MonoDETR — MonoDETR: Depth-guided Transformer for Monocular 3D Object Detection (ICCV 2023) · better on 2 of 2 shared results
- MonoMAE — MonoMAE: Enhancing Monocular 3D Detection through Depth-Aware Masked Autoencoders (NeurIPS 2024) · better on 0 of 1 shared result
- MonoUNI — MonoUNI: A Unified Vehicle and Infrastructure-side Monocular 3D Object Detection Network with Sufficient Depth Clues (NeurIPS 2023) · better on 0 of 1 shared result
- DEVIANT — DEVIANT: Depth EquiVarIAnt NeTwork for Monocular 3D Object Detection (ECCV 2022) · better on 1 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
- Rethinking Geometric Depth in Monocular 3D Object Detection: A Projection-Consistent Reformulation (NeurIPS 2026 · similarity 0.91)
- Learning Auxiliary Monocular Contexts Helps Monocular 3D Object Detection (AAAI 2022 · similarity 0.90)
- MonoCD: Monocular 3D Object Detection with Complementary Depths (CVPR 2024 · similarity 0.90)
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.90)
- MonoMAE: Enhancing Monocular 3D Detection through Depth-Aware Masked Autoencoders (NeurIPS 2024 · similarity 0.89)
- MonoDiff: Monocular 3D Object Detection and Pose Estimation with Diffusion Models (CVPR 2024 · similarity 0.89)
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.
