Research map / 3D vision / 3D generation
Weakly but Deeply Supervised Occlusion-Reasoned Parametric Road Layouts
Buyu Liu, Bingbing Zhuang, Manmohan Chandraker
Check 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 PS to the RGB-to-TPP model (ablation) | RGB road-layout model · KITTI | F1: .176 -> .425 | Improved |
| Add perspective-to-top-view transformation after PS (ablation) | RGB+PS · KITTI | F1: .425 -> 441 | Improved |
| Add the top-view hallucination module after transformation (ablation) | RGB+PS+T · KITTI | F1: 441 -> .473 | Improved |
| Add PS to the RGB-to-TPP model (ablation) | RGB road-layout model · KITTI | MSE: .230 -> .159 | Improved |
| Add perspective-to-top-view transformation after PS (ablation) | RGB+PS · KITTI | MSE: .159 -> .144 | Improved |
| Add the top-view hallucination module after transformation (ablation) | RGB+PS+T · KITTI | MSE: .144 -> .140 | Improved |
| Add PS to the RGB-to-TPP model (ablation) | RGB road-layout model · KITTI | Accu.-Mc.: .778 -> .827 | Improved |
| Add PS and top-view transformation (ablation) | RGB road-layout model · KITTI | Accu.-Bi.: .811 -> .833 | Improved |
| Use the full PS, transformation, and hallucination design instead of RGB-only (ablation) | Road-layout prediction · KITTI | F1: .176 -> .473 | Improved |
| Use the full method rather than RGB baseline | Road-layout prediction · NuScenes | F1: .109 -> .130 | Improved |
| Jointly train the full model end-to-end after module pretraining | Full model · Experiments | performance: 1% | Improved |
Closest papers by meaning
- Learning to Look around Objects for Top-View Representations of Outdoor Scenes (ECCV 2018 · similarity 0.91)
- A Parametric Top-View Representation of Complex Road Scenes (CVPR 2019 · similarity 0.89)
- RoadSceneBench: A Lightweight Benchmark for Mid-Level Road Scene Understanding (CVPR 2026 · similarity 0.87)
- Predicting Semantic Map Representations From Images Using Pyramid Occupancy Networks (CVPR 2020 · similarity 0.86)
- Understanding Road Layout From Videos as a Whole (CVPR 2020 · similarity 0.86)
- Achieving Speed-Accuracy Balance in Vision-based 3D Occupancy Prediction via Geometric-Semantic Disentanglement (AAAI 2025 · similarity 0.86)
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.
