fudan-zvg / RoadNet

[ICCV2023 Oral] RoadNetworkTRansformer & [AAAI 2024] LaneGraph2Seq
MIT License
68 stars 5 forks source link

Road Network and Lane Topology Extraction

ICCV 2023 (Oral) 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 Fudan University, Huawei Noah’s Ark Lab, Shanghai AI Lab

AAAI 2024 LaneGraph2Seq: Lane Topology Extraction with Language Model via Vertex-Edge Encoding and Connectivity Enhancement, Renyuan Peng, Xinyue Cai, Hang Xu, Jiachen Lu, Feng Wen, Wei Zhang, Li Zhang Fudan University, Huawei Noah’s Ark Lab

intro-fig

Release Announcement

Torch < 2.0 (Stable Version) Get Started

Please checkout for get_started.md The base path will be RoadNet/RoadNetwork-1.8.1/

Pre-training

./tools/dist_train.sh projects/configs/road_seg/lss_roadseg_48x32_b4x8_resnet_adam_24e.py 8

Move the final checkpoint to ckpts/

cp work_dirs/lss_roadseg_48x32_b4x8_resnet_adam_24e/latest.pth ckpts/lss_roadseg_48x32_b4x8_resnet_adam_24e.pth

Training

For ICCV: Translating Images to Road Network: A Non-Autoregressive Sequence-to-Sequence Approach

./tools/dist_train.sh projects/configs/rntr_ar_roadseq/lss_ar_rntr_l6_d256_gpu2x8.py 8

Torch >= 2.1 (Efficient Version) Get Started

Please checkout for get_started.md

Pre-training

./tools/dist_train.sh RoadNetwork/configs/road_seg/lss_roadseg_48x32_b4x8_resnet_adam_24e.py 8

Move the final checkpoint to ckpts/

cp work_dirs/lss_roadseg_48x32_b4x8_resnet_adam_24e/last_checkpoint ckpts/lss_roadseg_48x32_b4x8_resnet_adam_24e.pth

Training

For ICCV: Translating Images to Road Network: A Non-Autoregressive Sequence-to-Sequence Approach

./tools/dist_train.sh projects/RoadNetwork/configs/rntr_ar_roadseg/lss_ar_rntr_changeloss_test_fp16_torch2.py 8

For AAAI: LaneGraph2Seq: Lane Topology Extraction with Language Model via Vertex-Edge Encoding and Connectivity Enhancement

./tools/dist_train.sh projects/RoadNetwork/configs/lanegraph2seq/langraph2seq_fp16_torch2.py 8

Acknowledgements

We thank numerous excellent works and open-source codebases:

📜 BibTex

@inproceedings{lu2023translating,
  title={Translating Images to Road Network: A Non-Autoregressive Sequence-to-Sequence Approach},
  author={Lu, Jiachen and Peng, Renyuan and Cai, Xinyue and Xu, Hang and Li, Hongyang and Wen, Feng and Zhang, Wei and Zhang, Li},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={23--33},
  year={2023}
}
@inproceedings{peng2024lanegraph2seq,
  title={LaneGraph2Seq: Lane Topology Extraction with Language Model via Vertex-Edge Encoding and Connectivity Enhancement},
  author={Peng, Renyuan and Cai, Xinyue and Xu, Hang and Lu, Jiachen and Wen, Feng and Zhang, Wei and Zhang, Li},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year={2024}
}