BN-HTR_LS

This repository is based on the work done for the Bangla Handwritten Line Segmentation from Images.

Please do not use this code without proper citation.

For any further query, feel free to contact: [email protected]

*** INSTRUCTIONS to run the code ***

STEP-1 : Mount your Google Drive locally by using drive folder location.

STEP-2 : Set your own INPUT and OUTPUT directories.

STEP-3 : Run each section one after one.(Each execution results in one output only.)

STEP-4 : Set "plot_image = True" if you want to plot all images during execution.

=> See the test_data, Output folders to check the input, output sample.

Paper Information

@incollection{rahman2023bn,
 title={BN-HTRd: A Benchmark Dataset for Document Level Offline Bangla Handwritten Text Recognition (HTR) and Line Segmentation},
 author={Rahman, Md Ataur and Tabassum, Nazifa and Paul, Mitu and Pal, Riya and Islam, Mohammad Khairul},
 booktitle={Computer Vision and Image Analysis for Industry 4.0},
 pages={1--16},
 year={2023},
 publisher={CRC Press 6000 Broken Sound Parkway NW, Suite 300, Boca Raton, FL 33487-2742}
}

Dataset Information

 @misc{https://doi.org/10.17632/743k6dm543.3,
  doi = {10.17632/743K6DM543.3},
  url = {https://data.mendeley.com/datasets/743k6dm543/3},
  author = {{Md Ataur Rahman}},
  title = {BN-HTRd: A Benchmark Dataset for Document Level Offline Bangla Handwritten Text Recognition (HTR)},
  publisher = {Mendeley},
  year = {2022},
  copyright = {Creative Commons Attribution 4.0 International}
}