Research map / Graph learning / Graph self-supervised learning
TopDiG: Class-Agnostic Topological Directional Graph Extraction From Remote Sensing Images
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 TCND instead of FPN as the node detector (ablation) | TopDiG · Inria | mIoUtopo: 67.26 -> 68.39 | Helped |
| Add the TCND weighting block (ablation) | TopDiG · Inria | mIoUtopo: 43.74% to 68.39% | Helped |
| Set detected node count N to 320 (ablation) | TopDiG · Inria | mIoUtopo: 68.39 % | Helped |
| Increase node distance tolerance Φ from 0 to 2 (ablation) | TopDiG · Inria | mIoUtopo: 68.15 % to 68.39 % | Helped |
| Use TopDiG rather than DeepLabV3 | remote-sensing graph extraction · Inria | APLS: 40.48 -> 48.09 | Helped |
| Use TopDiG rather than PolyWorld | remote-sensing graph extraction · CrowdAI | APLS: 51.73 -> 59.61 | Helped |
| Use TopDiG rather than DeepLabV3 | remote-sensing graph extraction · GID | mIoUtopo: 70.55 -> 74.51 | Helped |
| Use TopDiG rather than DeepLabV3 | remote-sensing graph extraction · GF2 | mIoUtopo: 70.70 -> 75.28 | Helped |
| Use TopDiG rather than DeepLabV3 | road graph extraction · Massachusetts | APLS: 51.40 -> 64.60 | Helped |
Closest papers by meaning
- Towards Satellite Image Road Graph Extraction: A Global-Scale Dataset and A Novel Method (CVPR 2025 · similarity 0.85)
- Topograph: An Efficient Graph-Based Framework for Strictly Topology Preserving Image Segmentation (ICLR 2025 · similarity 0.84)
- Graph Out-of-Distribution Detection Goes Neighborhood Shaping (ICML 2024 · similarity 0.83)
- Structure-Adaptive Multi-View Graph Clustering for Remote Sensing Data (AAAI 2025 · similarity 0.81)
- Linear Transformer Topological Masking with Graph Random Features (ICLR 2025 · similarity 0.81)
- Point-Level Topological Representation Learning on Point Clouds (ICML 2025 · similarity 0.81)
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.