Research map / Label-efficient and robust learning / Recognition and long-tail
DeepRoadMapper: Extracting Road Topology From Aerial Images
Gellert Mattyus, Wenjie Luo, Raquel Urtasun
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 |
|---|---|---|---|
| use soft IoU loss instead of cross-entropy (ablation) | segmentation network · TorontoCity segmentation (validation) | mean IoU: 71.6 -> 75.6 | Improved |
| use soft IoU loss instead of cross-entropy (ablation) | segmentation network · TorontoCity segmentation (validation) | road IoU: 69.9 -> 75.1 | Improved |
| add connectivity reasoning (shortest-path gap filling) to basic centerline extraction (ablation) | Ours basic · TorontoCity road topology (test) | topology F1: 78.0 -> 83.9 | Improved |
| add connectivity reasoning to basic centerline extraction (ablation) | Ours basic · TorontoCity road topology (test) | topology precision: 71.2 -> 84.4 | Improved |
| add connectivity reasoning to basic centerline extraction (ablation) | Ours basic · TorontoCity road topology (test) | topology recall: 86.2 -> 83.3 | Hurt |
| add the Inception connection classifier (ablation) | Ours - NoDeepCon. · TorontoCity road topology (test) | topology F1: 83.9 -> 84.0 | Improved |
| use our deep segmentation as unary potentials in the baseline MRF | Wenger et al. [25] · TorontoCity road topology (test) | topology F1: 39.7 -> 63.1 | Improved |
| replace baseline [25] with our full method | Wenger et al. [25] · TorontoCity road topology (test) | topology precision: 26.1 -> 84.5 | Improved |
Closest papers by meaning
- TopoAL: An Adversarial Learning Approach for Topology-Aware Road Segmentation (ECCV 2020 · similarity 0.90)
- RoadTracer: Automatic Extraction of Road Networks From Aerial Images (CVPR 2018 · similarity 0.90)
- Improved Road Connectivity by Joint Learning of Orientation and Segmentation (CVPR 2019 · similarity 0.88)
- Topology Preserving Local Road Network Estimation From Single Onboard Camera Image (CVPR 2022 · similarity 0.87)
- DeepDualMapper: A Gated Fusion Network for Automatic Map Extraction Using Aerial Images and Trajectories (AAAI 2020 · similarity 0.87)
- VectorMapNet: End-to-end Vectorized HD Map Learning (ICML 2023 · similarity 0.87)
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.
