Abstract
Optical character recognition (OCR) involves the electronic transcription of handwritten, printed text from either scanned documents or any sort of images. This multiclass classification problem comprising of recognizing the various characters in a language and correctly classifying them has found its way in plenty of computer vision applications. Although much work has been done for the English language, there have been only a few explorations pertaining to OCR for the Devanagari script. This paper proposes a hybrid CNN-RNN model to classify characters using the Devanagari handwritten character dataset. The main objective is to design a model with higher accuracy than the CNN model reported in literature for the same purpose. The models are trained and evaluated using the same procedures. On evaluating the models, the hybrid CNN-RNN model has a testing accuracy of 98.71%, which is higher than the CNN model, having 97.71% testing accuracy. It also fares better than the standard neural network architectures-VGG16 and AlexNet which when taken without the pre-trained weights result in 97.62 and 98.20% testing accuracy respectively. Hence, this successfully demonstrates the attributes of RNN in improved feature extraction when used along with convolutional layers.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Mori S, Suen C, Yamamoto K (1992) Historical review of OCR research and development. Proc IEEE 80:1029–1058. https://doi.org/10.1109/5.156468
Bahdanau D, Cho K, Bengio Y (2014) Neural machine translation by jointly learning to align and translate. ICLR arXiv:1409.0473
Ruck D, Rogers S, Kabrisky M, Mills J (1990) Multisensor fusion classification with a multilayer perceptron. In: 1990 IJCNN international joint conference on neural networks. IEEE, pp 863–868. https://doi.org/10.1109/ijcnn.1990.137802
Lee H, Grosse R, Ranganath R, Ng AY (2009) Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations. In: Proceedings of the 26th annual international conference on machine learning-ICML’09. ICML, pp 609–616. https://doi.org/10.1145/1553374.1553453
Lecun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proc IEEE 86:2278–2324. https://doi.org/10.1109/5.726791
Guo Y, Liu Y, Bakker EM et al (2017) CNN-RNN: a large-scale hierarchical image classification framework. Multimedia Tools Appl 77:10251–10271. https://doi.org/10.1007/s11042-017-5443-x
Yao H, Zhang X, Zhou X, Liu S (2019) Parallel structure deep neural network using CNN and RNN with an attention mechanism for breast cancer histology image classification. Cancers 11:1901. https://doi.org/10.3390/cancers11121901
Acharya S, Pant AK, Gyawali PK (2015) Deep learning based large scale handwritten Devanagari character recognition. In: 2015 9th International conference on software, knowledge, information management and applications (SKIMA). IEEE, pp 1–6. https://doi.org/10.1109/skima.2015.7400041
Fischer SR (2004) History of writing. Reaktion Books, London, pp 32–37
Kavya A, Vivek N, Harika M, Nidumolu V (2020) Handwritten Devanagari character classification using CNN. Lect Notes Networks Syst 145:309–317. https://doi.org/10.1007/978-981-15-7345-3_25
Jain M, Mathew M, Jawahar C (2017) Unconstrained OCR for Urdu using deep CNN-RNN hybrid networks. In: 2017 4th IAPR Asian conference on pattern recognition (ACPR). IEEE, pp 747–752. https://doi.org/10.1109/acpr.2017.5
Deng Y, Kanervisto A, Rush A (2016) What you get is what you see: a visual markup decompiler. In: Association for the advancement of artificial intelligence. arXiv, abs/1609.04938, pp 32–37
Mishra V, Kaur D (2018) Sequence-to-sequence learning using deep learning for optical character recognition (OCR). In: 2018 International conference on computational science and computational intelligence (CSCI). IEEE, pp 324–329. https://doi.org/10.1109/csci46756.2018.00069
Fukushima K (1980) Neocognitron: a self-organising neural network model for a mechanism of pattern recognition unaffected by shift in position. Biol Cybern 36:193–202. https://doi.org/10.1007/bf00344251
Sherstinsky A (2020) Fundamentals of recurrent neural network (RNN) and long short-term memory (LSTM) network. In: Special Issue on machine learning and dynamical systems. J Physica D 404:132306. Elsevier. https://doi.org/10.1016/j.physd.2019.132306
Park S, Kwak N (2017) Analysis on the dropout effect in convolutional neural networks. In: Computer vision-ACCV 2016. Springer, Cham, pp 189–204. https://doi.org/10.1007/978-3-319-54184-6_12
Krizhevsky A, Sutskever I, Hinton G (2017) Imagenet classification with deep convolutional neural networks. Commun ACM 60:84–90. https://doi.org/10.1145/3065386
Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. ICLR arXiv:1409.1556
Paredes-Valverde M, Colomo-Palacios R, Salas-Zárate M, Valencia-GarcÃa R (2017) Sentiment analysis in spanish for improvement of products and services: a deep learning approach. Sci Program 2017:1–6. https://doi.org/10.1155/2017/1329281
Salas-Zárate, MDP, Valencia-GarcÃa R, Ruiz-MartÃnez A, Colomo-Palacios R (2016) Feature-based opinion mining in financial news: an ontology-driven approach. J Inf Sci43:458–479. https://doi.org/10.1177/0165551516645528
Prechelt L (1998) Automatic early stopping using cross validation: quantifying the criteria. Neural Netw 11:761–767. https://doi.org/10.1016/s0893-6080(98)00010-0
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Sansowa, R., Abraham, V., Patel, M.I., Gajjar, R. (2023). OCR for Devanagari Script Using a Deep Hybrid CNN-RNN Network. In: Dhavse, R., Kumar, V., Monteleone, S. (eds) Emerging Technology Trends in Electronics, Communication and Networking. Lecture Notes in Electrical Engineering, vol 952. Springer, Singapore. https://doi.org/10.1007/978-981-19-6737-5_22
Download citation
DOI: https://doi.org/10.1007/978-981-19-6737-5_22
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-19-6736-8
Online ISBN: 978-981-19-6737-5
eBook Packages: Computer ScienceComputer Science (R0)