Deep learning-based isolated handwritten Sindhi character recognition

Asghar Ali Chandio; Mehwish Leghari; Ali Orangzeb Panhwar; Shah Zaman Nizamani; Mehjabeen Leghari

doi:10.17485/IJST/v13i25.914

Article

Deep learning-based isolated handwritten Sindhi character recognition

VIEWS 2063
PDF 268

Indian Journal of Science and Technology

DOI: 10.17485/IJST/v13i25.914

Year: 2020, Volume: 13, Issue: 25, Pages: 2565-2574

Original Article

Deep learning-based isolated handwritten Sindhi character recognition

Asghar Ali Chandio¹,^2*, Mehwish Leghari²,³, Ali Orangzeb Panhwar⁴, Shah Zaman Nizamani², Mehjabeen Leghari³

¹School of Engineering and Information Technology, University of New South Wales, Australia. Tel.: +92-3343138628
²Department of Information Technology Quaid-e-Awam University of Engineering, Science & Technology, Nawabshah, Pakistan
³Department of Information Technology, University of Sindh, Pakistan
⁴Benazir Bhutto Shaheed University, Lyari, Pakistan
*Corresponding Author
Email: [email protected]

Received Date:12 June 2020, Accepted Date:02 July 2020, Published Date:21 July 2020

This work is licensed under a Creative Commons Attribution 4.0 International License.

Abstract

Motivation : The problem of handwritten text recognition is vastly studied since last few decades. Many innovative ideas have been developed, where state-of-the-art accuracy is achieved for the English, Chinese or Indian scripts.The recent developments for the cursive scripts such as Arabic and Urdu handwritten text recognition have achieved remarkable accuracy. However, for the Sindhi script, existing systems have not shown significant results and the problem is still an open challenge. Several challenges such as variations in writing styles, joined text, ligature overlapping, and others associated to the handwritten Sindhi text make the problem more complex. Objectives: In this study, a deep residual network with shortcut connections and summation fusion method using convolutional neural network (CNN) is proposed for automatic feature extraction and classification of handwritten Sindhi characters. Method: To increase the powerful feature representation ability of the network, the features of the convolutional layers in the residual block are fused together and combined with the output of the previous residual block. The proposed network is trained on a custom developed handwritten Sindhi character dataset. To tackle the problem of small data, a data augmentation with rotation, flipping and image enhancement techniques have been used. Findings: The experimental results show that the proposed model outperforms than the best results previously published for the handwritten Sindhi character recognition. Novelty: This is the first research that proposes deep residual network with summation fusion for the Sindhi handwritten text recognition.

Keywords: Handwritten Sindhi character recognition; Sindhi text recognition; cursive text recognition; deep learning; ResNet; convolutional neural network

References

Leghari M, Rahman MU. Towards Transliteration between Sindhi Scripts Using Roman Script. Linguistics and Literature Review. 2015;1(2):101–110. Available from: https://dx.doi.org/10.32350/llr.12.03
Hakro DN, Talib AZ. Printed Text Image Database for Sindhi OCR. ACM Transactions on Asian and Low-Resource Language Information Processing. 2016;15(4):1–18. Available from: https://dx.doi.org/10.1145/2846093
Hakro DN, Ismaili IA, Talib AZ, Bhatti Z, Mojai GN. Issues and challenges in Sindhi OCR. Sindh University Research Journal-SURJ. 2014;46(2):143–152.
Boufenar C, Kerboua A, Batouche M. Investigation on deep learning for off-line handwritten Arabic character recognition. Cognitive Systems Research. 2018;50:180–195. Available from: https://dx.doi.org/10.1016/j.cogsys.2017.11.002
Ahmed SB, Naz S, Swati S, Razzak MI. Handwritten Urdu character recognition using one-dimensional BLSTM classifier. Neural Computing and Applications. 2019;31(4):1143–51.
Ghanim TM, Khalil MI, Abbas HM. Comparative Study on Deep Convolution Neural Networks DCNN-Based Offline Arabic Handwriting Recognition. IEEE Access. 2020;8:95465–95482. Available from: https://dx.doi.org/10.1109/access.2020.2994290
Mathew M, Singh AK, Jawahar CV. Multilingual OCR for indic scripts. proceedings of 12th IEEE IAPR Workshop on Document Analysis Systems (DAS). 2016;p. 186–191.
Islam N, Islam Z, Noor N. 2017. Available from: https://arxiv.org/ftp/arxiv/papers/1710/1710.05703.pdf
Awan SA, Abro ZH, Jalbani AH, Hameed M. Handwritten Sindhi Character Recognition Using Neural Networks. Mehran University Research Journal of Engineering and Technology. 2018;37(1):6.
Nizamani AM, Janjua NU. Isolated Handwritten Character Recognition in Sindhi Language using Artificial Neural Network. Journal of Independent Studies and Research. 2012;10(1).
Kumari A, Sangrasi DM, Bhatti S, Chowdhry BS, Kumari S. Off-line Sindhi Handwritten Character Identification. International Journal of Information Technology and Computer Science. 2019;11(6):9–17. Available from: https://dx.doi.org/10.5815/ijitcs.2019.06.02
Shaikh NA, Mallah GA, Shaikh ZA. Character segmentation of Sindhi, an Arabic style scripting language, using height profile vector. Australian Journal of Basic and Applied Sciences. 2009;3(4):4160–4169.
Memon NA, Abbasi F, Zardari S. Glyph Identification and Character Recognition for Sindhi OCR. Mehran University Research Journal of Engineering and Technology. 2017;36(4).
Sanjrani AA, Baber J, Bakhtyar M, Noor W, Khalid M. Handwritten optical character recognition system for Sindhi numerals. Proceedings of IEEE International Conference on Computing, Electronic and Electrical Engineering. 2016;p. 262–267.
Ali I, Ali I, Subhash AK, Raza SA, Hassan B, Bhatti P. 2019. Available from: http://paper.ijcsns.org/07_book/201905/20190526.pdf
Solangi YA, Solangi ZA, Raza A, Shaikh NA, Mallah GA, Shah A. Offline-printed Sindhi Optical Text Recognition: Survey. Proceedings of IEEE 5th International Conference on Engineering Technologies and Applied Sciences. 2018;p. 1–5.
Chandio AA, Leghari M, Hakro D, Awan S, Jalbani AH. A Novel Approach for Online Sindhi Handwritten Word Recognition using Neural Network. Sindh University Research Journal-SURJ. 2016;48(1):213–216.
Chandio AA, Jalbani AH, Laghari M, Awan SA. Multi-Digit Handwritten Sindhi Numerals Recognition using SOM Neural Network. Mehran University Research Journal of Engineering and Technology. 2017;36(4).
Chandio AA, Leghari M, Leghari M, Jalbani AH. Multi-Font and Multi-Size Printed Sindhi Character Recognition using Convolutional Neural Networks. Pakistan Journal of Engineering and Applied Sciences. 2019;24(1):36–42.
He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition. 2016;p. 770–778.
He K, Zhang X, Ren S, Sun J. Identity mappings in deep residual networks. In: Proceedings of European conference on computer vision. (pp. 630-645) 2016.

Copyright

© 2020 Chandio, Leghari, Orangzeb Panhwar, Zaman Nizamani, Leghari.This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Published By Indian Society for Education and Environment (iSee)