Handwritten Dzongkha Alphabet Recognition System using Convolutional Neural Network
Keywords:
OCR, Deep Learning, CNN, Pattern recognition, Dzongkha languageAbstract
Pattern recognition is one of the fields in computer vision. With the advancement of deep learning technology, many machine learning algorithms were deployed for classification problems. Optical Character Recognition (OCR) is a method of processing and recognizing a character from a handwritten character or a printed document within a digital image. In this paper, an implementation using Convolutional Neural Network (CNN) was proposed for the classification of Handwritten Dzongkha alphabets. The dataset consists of 30 classes, each representing an alphabet of the Dzongkha language with 500 images in each class amounting to a total of 15000 images. Four layered CNN with a kernel size of 3 produced the optimal result in building the model and achieved an accuracy of 97.22% and a loss of 17.62%. This research is carried out for the first time in Bhutan and the findings from the study will act as the benchmark for future researchers. In the future, more handwritten alphabets need to be collected and trained with pre-trained models to get better accuracy.
References
Vidushi and M. Agarwal, “Intelligent Handwritten Digit Recognition Based on Multiple Parameters using CNN,” International Journal of Computer Sciences and Engineering, Vol.7, Issue.5, pp.636–641, 2019.
M. S. K. Dasari, “Optical Character Recognition of Devanagari Script Using Machine Learning- A Survey,” Journal of Xi’an University of Architecture & Technology, Vol. 12, Issue.8, pp.593–599, 2020.
D. Yadav, S. Sanchez-Cuadrado, and J. Morato, “Optical Character Recognition for Hindi Language Using a Neural-network Approach,” J. Inf. Process. Syst., Vol.9, Issue.1, pp. 117–140, 2013.
M. Avadesh and N. Goyal, “Optical Character Recognition for Sanskrit Using Convolution Neural Networks,” in 2018 13th IAPR International Workshop on Document Analysis Systems (DAS), pp. 447–452, 2018.
H. Yadav and S. Jain, “Offline Handwritten Character Recognition using Neural Networks,” International Journal of Computer Sciences and Engineering, Vol.7, Issue.5, pp.838–845, 2019.
Y. A. Nanehkaran, D. Zhang, S. Salimi, J. Chen, Y. Tian, and N. Al-Nabhan, “Analysis and comparison of machine learning classifiers and deep neural networks techniques for recognition of Farsi handwritten digits,” J. Supercomput., Vol.77, Issue.4, pp.3193–3222, 2021.
Y. Bengio, “Practical Recommendations for Gradient-Based Training of Deep Architectures,” in Neural Networks: Tricks of the Trade: Second Edition, G. Montavon, G. B. Orr, and K.-R. Müller, Eds. Berlin, Heidelberg: Springer, pp.437–478, 2012.
N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov, “Dropout: A Simple Way to Prevent Neural Networks from Overfitting,” J. Mach. Learn. Res., Vol.15, Issue.56, pp.1929–1958, 2014.
O. K. Oyedotun, A. E. R. Shabayek, D. Aouada, and B. Ottersten, “Going Deeper With Neural Networks Without Skip Connections,” in 2020 IEEE International Conference on Image Processing (ICIP), pp.1756–1760, 2020.
D. P. Kingma and J. Ba, “Adam: A Method for Stochastic Optimization,” ArXiv14126980 Cs, Jan. 2017, Accessed: Aug. 13, 2021. [Online]. Available: http://arxiv.org/abs/1412.6980
R. Poojary and A. Pai, “Comparative Study of Model Optimization Techniques in Fine-Tuned CNN Models,” in 2019 International Conference on Electrical and Computing Technologies and Applications (ICECTA), pp.1–4, 2019.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors contributing to this journal agree to publish their articles under the Creative Commons Attribution 4.0 International License, allowing third parties to share their work (copy, distribute, transmit) and to adapt it, under the condition that the authors are given credit and that in the event of reuse or distribution, the terms of this license are made clear.
