Research map / Detection and segmentation / 3D detection and driving
Learning And-Or Model to Represent Context and Occlusion for Car Detection and Viewpoint Estimation
Tianfu Wu, Bo Hu Li, Song Chun Zhu
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 multi-car context and occlusion modeling over DPM | Hierarchical And-Or model · KITTI (split trainset) | AP: outperforms DPM by 9.1% | Improved |
| add multi-car context and occlusion modeling over DPM | Hierarchical And-Or model · Parking Lot | AP: 55.2% AP, outperforms DPM by 10.9% | Improved |
| add context to occlusion-only And-Or Structure (ablation) | Hierarchical And-Or model · Street Parking | AP: 57.8% -> 62.1% | Improved |
| replace greedy part selection with CAD-simulated semantic parts (ablation) | AOG hierarchical model · Street Parking | AP: 62.1% -> 65.3% | Improved |
| train on full KITTI training set instead of half | AOG+Greedy · KITTI test benchmark (Moderate) | AP: 71.88% -> 75.94% | Improved |
| train on full KITTI training set instead of half | AOG+Greedy · KITTI test benchmark (Easy) | AP: 84.36% -> 84.80% | Improved |
| train on full KITTI training set instead of half | AOG+Greedy · KITTI test benchmark (Hard) | AP: 59.27% -> 60.70% | Improved |
| use And-Or Structure with CAD-learned occlusion over DPM | And-Or Structure · synthetic occlusion dataset | AP: outperforms DPM by 7.2% in AP | Improved |
| use And-Or Structure over DPM | And-Or Structure · PASCAL VOC2007 car | AP: 58.2% -> 58.7% | No clear effect |
| use hierarchical And-Or model over DPM | AOG+Greedy · PASCAL VOC2007 car | AP: 58.2% -> 60.6% | Improved |
| use AOG over DPM on KITTI moderate/hard | Hierarchical And-Or model · KITTI test (Moderate, Hard) | AP: above 10% in AP | Improved |
| use And-Or Structure for viewpoint estimation instead of VDPM | And-Or Structure · PASCAL3D+ car (4 views) | AVP: 20.2% -> 34.3% | Improved |
Closest papers by meaning
- Part-level Car Parsing and Reconstruction in Single Street View Images (TPAMI 2021 · similarity 0.88)
- Recognizing Car Fluents From Video (CVPR 2016 · similarity 0.87)
- Structural Information Guided Multimodal Pre-training for Vehicle-Centric Perception (AAAI 2024 · similarity 0.87)
- CorrBEV: Multi-View 3D Object Detection by Correlation Learning with Multi-modal Prototypes (CVPR 2025 · similarity 0.87)
- Infer the Whole from a Glimpse of a Part: Keypoint-Based Knowledge Graph for Vehicle Re-Identification (AAAI 2025 · similarity 0.85)
- Weakly Supervised Semantic Segmentation for Driving Scenes (AAAI 2024 · similarity 0.84)
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.
