Research map / Graph learning / Graph self-supervised learning
HDMapGen: A Hierarchical Graph Generative Model of High Definition Maps
Check 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 hierarchical graph representation instead of a sequence or plain graph | HD map generation · Argoverse | generation quality across urban planning, geometry, and topology distances: HDMapGen topology-first: 0.26, 4.06, 4.47, 9.9, 1.49, 0.37, 0.05, 0.63; PlainGen coordinate-first: 0.39, 4.30, 4.81, 12. | Helped |
| Use hierarchical graph generation rather than sequence generation | HD map generation · Argoverse, FoV 200m × 200m | generation time [s]: 2.28 -> 0.20 | Helped |
| Represent a map with a hierarchical graph rather than a plain graph | HD map generation · Argoverse, FoV 200m × 200m | generation time [s]: 0.89 -> 0.20 | Helped |
| Generate coordinates before topology (ablation) | HDMapGen global graph decoder · Argoverse | global graph BCE: 0.050 -> 0.001 | Helped |
| Generate topology before coordinates (ablation) | HDMapGen global graph decoder · Argoverse | global graph NLL: -5.610 -> -6.142 | Helped |
| Model coordinate-topology dependence instead of generating both independently (ablation) | HDMapGen global graph decoder · Argoverse | global graph NLL: -4.490 -> -6.142 | Helped |
| Model coordinate-topology dependence instead of generating both independently (ablation) | HDMapGen global graph decoder · Argoverse | global graph BCE: 0.053 -> 0.001 | Helped |
| Increase sampling temperature τ (ablation) | HDMapGen · In-house dataset | internal diversity (Chamfer distance × 10^4): 2.75 -> 8.34 | Helped |
| Increase sampling temperature τ (ablation) | HDMapGen · In-house dataset | generation realism | Hurt |
Closest papers by meaning
- Unveiling the Hidden: Online Vectorized HD Map Construction with Clip-Level Token Interaction and Propagation (NeurIPS 2024 · similarity 0.89)
- Driving by the Rules: A Benchmark for Integrating Traffic Sign Regulations into Vectorized HD Map (CVPR 2025 · similarity 0.88)
- Leveraging SD Map to Augment HD Map-based Trajectory Prediction (CVPR 2025 · similarity 0.88)
- T2SG: Traffic Topology Scene Graph for Topology Reasoning in Autonomous Driving (CVPR 2025 · similarity 0.87)
- MAPLM: A Real-World Large-Scale Vision-Language Benchmark for Map and Traffic Scene Understanding (CVPR 2024 · similarity 0.87)
- InteractionMap: Improving Online Vectorized HDMap Construction with Interaction (CVPR 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.