HMBD Dataset v1

HMBD v1 is an Arabic Handwritten Characters Dataset.

(1) Introduction:

The HMBD v1 dataset captures the different positions of the Arabic handwritten characters; isolated, beginning, middle, and end; besides, the numbers.

(2) Published Papers:

The HMBD v1 dataset is published in "A new Arabic handwritten character recognition deep learning system (AHCR-DLS)" where the construction, pre-processing, and compilation phases are discussed. Link: https://link.springer.com/article/10.1007/s00521-020-05397-2 DOI: https://doi.org/10.1007/s00521-020-05397-2

(3) Dataset Specifications:

• Version: 1.0.
• The number of classes (categories) is 115.
• The number of unique images is 54,115.
• The number of volunteers is 125.
• Each image dimension is 300 x 300 (i.e. width = 300 and height = 300).
• Background color: White.
• Character color: Black.

(4) Dataset Template:

The seven-page dataset template file used in collecting the dataset is stored in "Dataset Template v1.pdf".

(5) Directory Hierarchy:

The hierarchy of the folder is stored in "tree.txt" and "folders.txt". The first contains the folders' and files' names while the latter one contains only the folders' names.

(6) Citation:

Balaha, H.M., Ali, H.A., Saraya, M. et al. A new Arabic handwritten character recognition deep learning system (AHCR-DLS). Neural Comput & Applic 33, 6325–6367 (2021). https://doi.org/10.1007/s00521-020-05397-2

@article{balaha2021new,
  title={A new Arabic handwritten character recognition deep learning system (AHCR-DLS)},
  author={Balaha, Hossam Magdy and Ali, Hesham Arafat and Saraya, Mohamed and Badawy, Mahmoud},
  journal={Neural Computing and Applications},
  volume={33},
  number={11},
  pages={6325--6367},
  year={2021},
  publisher={Springer}
}

(7) Licence:

The HMBD dataset is licensed by CC BY-NC-SA 4.0.

The CC BY-NC-SA 4.0 is one of the Creative Commons (CC) licenses and allows the different users to share the script only if they (1) give the credits to the copyright holders, (2) do not use the script for any commercial purposes, and (3) distribute any additions, transformations or changes to the script under this same license.

Full Description: https://creativecommons.org/licenses/by-nc-sa/4.0/

(8) More Information About Me:

Online CV: https://hossambalaha.github.io/