Research map / Detection and segmentation / 3D detection and driving
Toward Unified 3D Object Detection via Algorithm and Data Unification
Zhuoling Li, Xiaogang Xu, Ser-Nam Lim, Hengshuang Zhao
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 proposal head (ablation) | UniMODE · Omni3D | AP3D: 12.3% -> 15.9% | Improved |
| add uneven BEV grid (ablation) | UniMODE · Omni3D | AP3D: 15.9% -> 16.6% | Improved |
| add sparse BEV feature projection (ablation) | UniMODE · Omni3D | AP3D: 16.6% -> 16.6% | No clear effect |
| add unified domain alignment (ablation) | UniMODE · Omni3D | AP3D: 16.6% -> 17.4% | Improved |
| use uneven depth bin split instead of even (ablation) | UniMODE · Omni3D | AP3D: 17.4% -> 15.3% | Hurt |
| reduce BEV grid size from 1 m to 0.5 m (ablation) | UniMODE · Omni3D | AP3D: 17.4% -> 18.1% | Improved |
| increase BEV grid size from 1 m to 2 m (ablation) | UniMODE · Omni3D | AP3D: 17.4% -> 16.5% | Hurt |
| raise sparse projection threshold tau from 0 to 1e-3 (ablation) | UniMODE · Omni3D | AP3D: 17.4% -> 17.4% | No clear effect |
| raise sparse projection threshold tau to 1e-1 (ablation) | UniMODE · Omni3D | AP3D: 17.4% -> 4.7% | Hurt |
| add DALN instead of no domain adaptation (ablation) | UniMODE · ARKitScenes to SUN-RGBD | AP3D (SUN-RGBD): 12.3% -> 13.0% | Improved |
| add DALN instead of direct regression (ablation) | UniMODE · ARKitScenes in-domain | AP3D (ARKitScenes): 33.9% -> 35.0% | Improved |
| add MIC guidance losses (ablation) | MM-UniMODE · MM-Omni3D | AP3D: 23.8% -> 26.1% | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PETR — PETR: Position Embedding Transformation for Multi-View 3D Object Detection (ECCV 2022) · better on 2 of 2 shared results
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
- UniMODE: Unified Monocular 3D Object Detection (CVPR 2024 · similarity 0.95)
- BEVDistill: Cross-Modal BEV Distillation for Multi-View 3D Object Detection (ICLR 2023 · similarity 0.91)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 2025 · similarity 0.90)
- Instance-Aware Multi-Camera 3D Object Detection with Structural Priors Mining and Self-Boosting Learning (AAAI 2024 · similarity 0.89)
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.89)
- GeoBEV: Learning Geometric BEV Representation for Multi-view 3D Object Detection (AAAI 2025 · 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.
