Research map / Graph learning / Graph self-supervised learning
Learning and Aggregating Lane Graphs for Urban Automated Driving
Martin Büchner, Jannik Zürn, Ion-George Todoran, Abhinav Valada, Wolfram Burgard
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 LaneGNN instead of skeletonized regression | Successor-LGP · UrbanLaneGraph | SDA20: 0.020 -> 0.227 | Improved |
| Use LaneGNN instead of LaneGraphNet | Successor-LGP · UrbanLaneGraph | TOPO P/R: 0.0/0.0 -> 0.600/0.699 | Improved |
| Use proposed iterative aggregation instead of naïve aggregation | LaneGNN · Full-LGP, Miami test set | Graph IoU: 0.376 -> 0.384 | Improved |
| Use proposed aggregation instead of LaneExtraction | Lane graph path planning · Miami test area | MMD [m]: 157.0 -> 2.2 | Improved |
| Use proposed aggregation instead of LaneExtraction | Lane graph path planning · Miami test area | SR: 0.47 -> 0.46 | Hurt |
| Add causal message passing (ablation) | LaneGNN · Successor-LGP, UrbanLaneGraph | APLS: 0.188 -> 0.202 | Improved |
| Add aerial edge features (ablation) | LaneGNN · Successor-LGP, UrbanLaneGraph | Graph IoU: 0.320 -> 0.347 | Improved |
| Use the ego-lane mask instead of the standard lane mask for proposal sampling (ablation) | LaneGNN · Successor-LGP, UrbanLaneGraph | Graph IoU: 0.250 -> 0.347 | Improved |
| Add aerial edge features rather than aerial node features (ablation) | LaneGNN · Successor-LGP, UrbanLaneGraph | Graph IoU: 0.320 -> 0.347 | Improved |
| Add the node loss term (ablation) | LaneGNN · Successor-LGP, UrbanLaneGraph | APLS: 0.144 -> 0.202 | Improved |
| Increase proposal nodes from 200 to 400 (ablation) | LaneGNN · Successor-LGP, UrbanLaneGraph | APLS: 0.077 -> 0.202 | Improved |
| Use six message-passing layers instead of none (ablation) | LaneGNN · Successor-LGP, UrbanLaneGraph | APLS: 0.057 -> 0.202 | Improved |
Closest papers by meaning
- Bezier Everywhere All at Once: Learning Drivable Lanes as Bezier Graphs (CVPR 2024 · similarity 0.90)
- LaneSegNet: Map Learning with Lane Segment Perception for Autonomous Driving (ICLR 2024 · similarity 0.90)
- LaneGraph2Seq: Lane Topology Extraction with Language Model via Vertex-Edge Encoding and Connectivity Enhancement (AAAI 2024 · similarity 0.88)
- GLane3D: Detecting Lanes with Graph of 3D Keypoints (CVPR 2025 · similarity 0.88)
- Sketch and Refine: Towards Fast and Accurate Lane Detection (AAAI 2024 · similarity 0.87)
- GEONet: Global Enhancement and Optimization Network for Lane Detection (AAAI 2025 · 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.
