CFP last date
21 October 2024
Reseach Article

Deep Arabic Font Family and Font Size Recognition

by Ibrahim M. Amer, Salma Hamdy, Mostafa G. M. Mostafa
International Journal of Computer Applications
Foundation of Computer Science (FCS), NY, USA
Volume 176 - Number 4
Year of Publication: 2017
Authors: Ibrahim M. Amer, Salma Hamdy, Mostafa G. M. Mostafa
10.5120/ijca2017915589

Ibrahim M. Amer, Salma Hamdy, Mostafa G. M. Mostafa . Deep Arabic Font Family and Font Size Recognition. International Journal of Computer Applications. 176, 4 ( Oct 2017), 1-6. DOI=10.5120/ijca2017915589

@article{ 10.5120/ijca2017915589,
author = { Ibrahim M. Amer, Salma Hamdy, Mostafa G. M. Mostafa },
title = { Deep Arabic Font Family and Font Size Recognition },
journal = { International Journal of Computer Applications },
issue_date = { Oct 2017 },
volume = { 176 },
number = { 4 },
month = { Oct },
year = { 2017 },
issn = { 0975-8887 },
pages = { 1-6 },
numpages = {9},
url = { https://ijcaonline.org/archives/volume176/number4/28537-2017915589/ },
doi = { 10.5120/ijca2017915589 },
publisher = {Foundation of Computer Science (FCS), NY, USA},
address = {New York, USA}
}
%0 Journal Article
%1 2024-02-07T00:41:35.605834+05:30
%A Ibrahim M. Amer
%A Salma Hamdy
%A Mostafa G. M. Mostafa
%T Deep Arabic Font Family and Font Size Recognition
%J International Journal of Computer Applications
%@ 0975-8887
%V 176
%N 4
%P 1-6
%D 2017
%I Foundation of Computer Science (FCS), NY, USA
Abstract

Font family and font size recognition became an essential step for document analysis. Font recognition helps to identify the proper segmentation method to be used before feeding the document to the Optical Character Recognition (OCR). In this paper, some of the previous techniques used for font family and font size recognition will be discussed then we will discuss the proposed method that is based on deep learning. Two methods have been presented in this paper 1) a method for font family recognition (font size invariant) and 2) a method for font size recognition. Both methods use Deep Convolutional Neural Networks (D-CNN). We evaluated the proposed method on Arabic Printed Text Image Database (APTI) [7] and on a document generated using APTI database word images and scanned with the scanner.

References
  1. Hamzah Luqman, Sabri A. Mahmoud and Sameh Awaida: Arabic and Farsi Font Recognition: Survey. International Journal of Pattern Recognition and Artificial Intelligence (IJPRAI). Vol. 29, No. 1 (2015).
  2. Slimane, Fouad, Slim Kanoun, Jean Hennebert, Adel M. Alimi, and Rolf Ingold. ”A study on font-family and font-size recognition applied to Arabic word images at ultra-low resolution.” Pattern Recognition Letters 34, no. 2 (2013): 209-218.
  3. Jaiem, F. K., Kanoun, S., & Eglin, V. (2014, September). Arabic font recognition based on a texture analysis. In Frontiers in Handwriting Recognition (ICFHR), 2014 14th International Conference on (pp. 673-677). IEEE.
  4. Mousa, Mahmoud AA, Mohammed S. Sayed, and Mahmoud I. Abdalla. ”An efficient algorithm for Arabic optical font recognition using scale-invariant detector.” International Journal on Document Analysis and Recognition (IJDAR) 18, no. 3 (2015): 263-270.
  5. Bozkurt, Alican, Pinar Duygulu, and A. Enis Cetin. ”Classifying fonts and calligraphy styles using complex wavelet transform.” Signal, Image and Video Processing 9, no. 1 (2015): 225-234.
  6. Ibrahim M.Amer, Salma Hamdy, Mostafa, M. G. Mostafa, ”Deep Arabic Document Layout Analysis” submitted to International conference on Intelligent Computing and Information Systems (ICICS). Aug. 2017.
  7. APTI Arabic Printed Text Image Database. https: //diuf.unifr.ch/diva/APTI/index.html
  8. Srivastava, Nitish, Geoffrey E. Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. ”Dropout: a simple way to prevent neural networks from overfitting.” Journal of Machine Learning Research 15, no. 1 (2014): 1929-1958.
  9. Kingma, Diederik, and Jimmy Ba. ”Adam: A method for stochastic optimization.” arXiv preprint arXiv:1412.6980 (2014).
  10. Nair, Vinod, and Geoffrey E. Hinton. ”Rectified linear units improve restricted boltzmann machines.” In Proceedings of the 27th international conference on machine learning (ICML-10), pp. 807-814. 2010.
Index Terms

Computer Science
Information Sciences

Keywords

Font Family Recognition Font Size Recognition Optical Character Recognition (OCR) Document Layout Analysis (DLA) Deep Learning Deep Convolutional Neural Network (D-CNN)