High-quality synthetic text data generation for Urdu Text Recognition

Released as a supplement of UTRNet: High-Resolution Urdu Text Recognition

• Create dataset using the following command python3 run.py --count 1000 --length 10 --max_length 100 --random --name_format 2 --height 128 --thread_count 8 --skew_angle 10 --random_skew --blur 2 --random_blur --salt_and_pepper 0.1 --distorsion 3 --distorsion_orientation 2 --background 3 --random_fit --random_resize --random_crop --random_shearx --random_margins --margins 5,5,5,5 --output_dir 1k_images

• Run python3 run.py --help for help

1. Download the UTRSet-Synth dataset
2. For more information & other resources, visit Project Webpage
3. Main codebase - UTRNet Repo

1. Based on the trdg library
2. The code (& the generated dataset) is for research purposes only and must not be used for any other purpose without the author's explicit permission.

If you use the code/model/dataset, please cite the following paper:

@article{rahman2023utrnet,
      title={UTRNet: High-Resolution Urdu Text Recognition In Printed Documents}, 
      author={Abdur Rahman and Arjun Ghosh and Chetan Arora},
      journal={arXiv preprint arXiv:2306.15782},
      year={2023},
      eprint={2306.15782},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      doi = {https://doi.org/10.48550/arXiv.2306.15782},
      url = {https://arxiv.org/abs/2306.15782}
}

. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License for Noncommercial (academic & research) purposes only and must not be used for any other purpose without the author's explicit permission.