Research map / Graph learning / GNN architectures
TD-Road: Top-Down Road Network Extraction with Holistic Graph Construction
Yang He, Ravi Garg, Amber Roy Chowdhury
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 neighbor-enhanced context to relation reasoning (ablation) | TD-Road · City-Scale | APLS: 62.17 -> 65.74 | Improved |
| add neighbor-enhanced context to relation reasoning (ablation) | TD-Road · SpaceNet | APLS: 60.91 -> 65.15 | Improved |
| add neighbor-enhanced context to relation reasoning (ablation) | TD-Road · City-Scale | TOPO F1: 72.83 -> 76.43 | Improved |
| add neighbor-enhanced context to relation reasoning (ablation) | TD-Road · SpaceNet | TOPO F1: 77.74 -> 81.15 | Improved |
| use road-mask weighted BCE instead of plain BCE (ablation) | TD-Road (naive) · City-Scale (naive relation module) | APLS: 60.32 -> 62.17 | Improved |
| use road-mask weighted BCE instead of plain BCE (ablation) | TD-Road (naive) · City-Scale (naive relation module) | TOPO F1: 60.32 -> 62.17 | Improved |
| use road-mask weighted BCE instead of focal loss (ablation) | TD-Road (naive) · City-Scale (naive relation module) | APLS: 47.95 -> 62.17 | Improved |
| use focal loss instead of plain BCE (ablation) | TD-Road (naive) · City-Scale (naive relation module) | precision: 75.71 -> 80.82 | Mixed |
| vary connectedness threshold 0.3 to 0.5 (ablation) | TD-Road · City-Scale | APLS | No clear effect |
| vary key point threshold 0.3 to 0.5 (ablation) | TD-Road · City-Scale | APLS | Hurt |
| replace ResNet-34 with DLA backbone (Sat2Graph) versus ours | Sat2Graph vs TD-Road · City-Scale | APLS: 63.14 vs 65.74 | Mixed |
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.
- Sat2Graph — Sat2Graph: Road Graph Extraction through Graph-Tensor Encoding (ECCV 2020) · better on 5 of 5 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
- LineGraph2Road: Structural Graph Reasoning on Line Graphs for Road Network Extraction (ECCV 2026 · similarity 0.91)
- Translating Images to Road Network: A Sequence-to-Sequence Perspective (TPAMI 2025 · similarity 0.89)
- Mastering Spatial Graph Prediction of Road Networks (ICCV 2023 · similarity 0.88)
- VecRoad: Point-Based Iterative Graph Exploration for Road Graphs Extraction (CVPR 2020 · similarity 0.88)
- Road Network Representation Learning with the Third Law of Geography (NeurIPS 2024 · similarity 0.87)
- LaneGraph2Seq: Lane Topology Extraction with Language Model via Vertex-Edge Encoding and Connectivity Enhancement (AAAI 2024 · 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.
