This is the official implementation of Paper: LRANet++: Low-Rank Approximation Network for Accurate and Efficient Text Spotting.
This implementation is based on mmocr-0.2.1, so please refer to it for detailed requirements. Our code has been tested with Pytorch-1.8.1 + cuda11.1 We recommend using Anaconda to manage environments. Run the following commands to install dependencies.
conda create -n lranet_pp python=3.7 -y
conda activate lranet_pp
conda install pytorch=1.8 torchvision cudatoolkit=11.1 -c pytorch -c nvidia -c conda-forge
pip install mmcv-full==1.3.9 -f https://download.openmmlab.com/mmcv/dist/cu111/torch1.8.0/index.html
pip install mmdet==2.15.1
git clone https://github.com/ychensu/LRANet-PP
cd LRANet-PP
pip install -r requirements.txt
python setup.py build develop
cd mmocr/models/textend2end/utils/grid_sample_batch
python setup.py build develop
- ICDAR2017-MLT [image]
- Syntext-150k: [image]
- ICDAR2015 [image]
- ICDAR2013 [image]
- CTW1500 [image]
- Total-Text [image]
The prepared annotations can be downloaded from Google Drive
Please download and extract the above datasets into the data folder following the file structure below.
data
├─totaltext
│ │ totaltext_train.json
│ │ totaltext_test.json
│ │ training
│ │ test
│
├─CTW1500
│ │ ctw1500_train.json
│ │ ctw1500_test.json
│ │ training
│ │ test
│
├─synthtext-150k
├─syntext1
│ │ syntext1_train.json
│ └─train_images
├─syntext2
│ syntext2_train.json
└─train_images
python orthanchors/generate_lra.py --json_path <your_json_path> --output_dir <output_dir>
CUDA_VISIBLE_DEVICES=0,1,2,3 bash tools/dist_train.sh configs/lranet_pp/lranet_pp_pretrain.py work_dirs/pretrain 4
CUDA_VISIBLE_DEVICES=0,1,2,3 bash tools/dist_train.sh configs/lranet_pp/lranet_pp_totaltext.py work_dirs/totaltext 4 --load-from work_dirs/pretrain/latest.pth
CUDA_VISIBLE_DEVICES=0 python tools/test.py configs/lranet_pp/lranet_pp_totaltext.py work_dirs/totaltext/latest.pth --eval hmean-e2e
Pretrain: OneDrive Total-Text : OneDrive CTW1500: OneDrive
@inproceedings{su2024lranet,
title={LRANet: Towards accurate and efficient scene text detection with low-rank approximation network},
author={Su, Yuchen and Chen, Zhineng and Shao, Zhiwen and Du, Yuning and Ji, Zhilong and Bai, Jinfeng and Zhou, Yong and Jiang, Yu-Gang},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
volume={38},
number={5},
pages={4979--4987},
year={2024}
}
@article{su2025lranet++,
title={LRANet++: Low-Rank Approximation Network for Accurate and Efficient Text Spotting},
author={Su, Yuchen and Chen, Zhineng and Du, Yongkun and Wu, Zuxuan and Xie, Hongtao and Jiang, Yu-Gang},
journal={arXiv preprint arXiv:2511.05818},
year={2025}
}