Research map / Architectures and efficiency / Attention and transformers
Translating Images to Road Network: A Non-Autoregressive Sequence-to-Sequence Approach
Jiachen Lu, Renyuan Peng, Xinyue Cai, Hang Xu, Hongyang Li, Feng Wen, Wei Zhang, Li Zhang
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 |
|---|---|---|---|
| decouple autoregressive dependency into semi-autoregressive generation | RoadNetTransformer · nuScenes multi-camera | reachability F1 / FPS: R-F 61.4 -> 67.1; FPS 0.1 -> 0.6 (6.0x) | Improved |
| use non-autoregressive masked iterative generation after SAR | RoadNetTransformer · nuScenes multi-camera | FPS: 0.1 -> 4.7 (47x); R-F 61.4 -> 66.0 | Improved |
| NAR instead of SAR (ablation) | RoadNetTransformer · nuScenes multi-camera | reachability F1: 67.1 -> 66.0 | Hurt |
| add Intra-seq self-attention in the parallel-seq decoder (ablation) | NAR-RNTR · nuScenes multi-camera | reachability F1: 56.3 -> 66.0 | Improved |
| increase key-point decoder layers from 3 to 6 (ablation) | NAR-RNTR · nuScenes multi-camera | landmark F1: 56.1 -> 60.2 | Improved |
| increase parallel-seq decoder layers from 3 to 6 (ablation) | NAR-RNTR · nuScenes multi-camera | landmark F1 / FPS: L-F 60.2 -> 61.3; FPS 4.7 -> 3.0 | Mixed |
| raise training mask ratio from 50% to 90% (ablation) | NAR-RNTR · nuScenes multi-camera | reachability F1: 51.3 -> 66.0 | Improved |
| increase inference iterations from 3 to 6 (ablation) | NAR-RNTR · nuScenes multi-camera | reachability F1: 66.0 -> 66.1; FPS 4.7 -> 2.8 | No clear effect |
| use RoadNet Sequence transformers instead of STSU/TPLR | front-camera road network extraction · nuScenes PON val | mean F1: 58.2 -> 59.3 (AR); 61.6 (SAR) | Improved |
Closest papers by meaning
- Translating Images to Road Network: A Sequence-to-Sequence Perspective (TPAMI 2025 · similarity 0.97)
- RoadTracer: Automatic Extraction of Road Networks From Aerial Images (CVPR 2018 · similarity 0.87)
- TD-Road: Top-Down Road Network Extraction with Holistic Graph Construction (ECCV 2022 · similarity 0.87)
- Dual-branch Spatial-Temporal Self-supervised Representation for Enhanced Road Network Learning (AAAI 2026 · similarity 0.87)
- Road Network Representation Learning with the Third Law of Geography (NeurIPS 2024 · similarity 0.85)
- Unveiling the Hidden: Online Vectorized HD Map Construction with Clip-Level Token Interaction and Propagation (NeurIPS 2024 · similarity 0.85)
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.
