Abstract
Handwritten character recognition is an important step toward automatic transliteration of the valuable historical documents archived in digital libraries. This is a challenging task as it requires a labor-intensive handcrafting of features from a huge set of character classes. Moreover, the intra-class variability of handwritten characters is high causing a major bottleneck for recognition accuracy. A deep convolutional neural networks (DCNN) approach for character recognition of handwritten historical Kannada manuscripts is presented in this paper. DCNN is a model that unifies feature extraction and classification. It inherently learns the most discriminative features from the given data, thus subverting the usage of handcrafted features. In this work, the features extracted from the characters using DCNN are fed to SGDM and SVM classification algorithms for recognition. This approach is experimented on the digitized estampages of historical Kannada stone inscriptions belonging to eleventh century and promising results are observed.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
Nunamaker, B., Bukahari, S.S., Borth, D., Dengel, A.: A tesseract-based OCR framework for Historical documents lacking groundtruth text. In: Proceedings of International Conference of Image Processing, pp. 2381–8549. IEEE (2016)
Chandrakala, H.T., Thippeswamy, G.: Epigraphic document image enhancement using retinex method. In: Proceedings of 3rd International Symposium of Signal Processing and Intelligent Recognition Systems, Thampi, S.M., Krishnan, S., Manuel, J., Rodriguez, C., Das, S., Wozniak, M., Al-Jumeily, D. (eds.) Advances in Signal Processing and Intelligent Recognition Systems. Springer International Publishing, ISBN: 978-3-319-67933-4 (2017)
Chandrakala, H.T., Thippeswamy, G.: A comprehensive survey on OCR techniques for kannada script. Int. J. Sci. Res. 5(4) (2016). ISSN: 2319-7064
Soumya, C.S., Thippeswamy, G., Srivani, P.: Design and development of rotation and scale invariant 3D models for efficient representation and accurate classification of human faces. Int. J. Adv. Electron. Comput. Sci. (2016). ISSN: 2393-2835
Amrutha Raj, V., Jyothi, R.L., AnilKumar, A.: Grantha script recognition from ancient palm leaves using histogram of orientation shape context. In: Proceedings of IEEE International Conference on Computing Methodologies and Communication (2017). 978-1-5090-4890-8
Cheriet, M., Moghaddam, R., Arabnejad, E., Zhong, G.: Manifold Learning for the Shape Based Recognition of Historical Arabic Documents, pp. 471–491. Elsevier (2013)
Vamvakas, G., Gatos, B., Perantonis, S.J.: A novel feature extraction and classification methodology for the recognition of historical documents. In: Proceedings of 10th International Conference on Document Analysis and Recognition (2009). 978-7695-3725-2
Richarz, J., Vajda, S., Grzeszick, R., Fink, G.A.: Semi supervised learning for character recognition in Historical archive documents. In: Pattern Recognition, vol. 47, pp. 1011–1020. Elsevier (2014)
Kim, M.S., Cho, K.T., Kwag, H.K., Kim, J.K.: Segmentation of handwritten characters for digitalizing Korean Historical documents. In: DAS, LNCS 3136, pp. 114–124, Springer (2004)
Bannigidada, P., Gudada, C.: Identification and classification of historical kannada handwritten document images using GLCM features. Int. J. Adv. Res. Comput. Sci. 4(1), 0976–5697 (2018)
Zha, S., Peng, X., Cao, H., Zhuang, X., Natarajan, P., Natarajan, P.: Text classification via iVector based feature representation. In: 11th IAPR International Workshop on Document Analysis Systems, IEEE (2014). 978-1-4799-3243-6
Soumya, A., Hemanth Kumar, G.: Recognition of historical records using gabor and zonal features. Signal Image Process. Int. J. 6(4), 57–69 (2015)
Vijayalakshmi, T.R.: Reduction of features to identify characters from degraded historical manuscripts. Alexandria Eng. J. 1110-0168, Elsevier (2017)
Zhang, G., Han, X.: Deep learning based tangut character recognition. In: Proceedings of 4th International Conference on Systems and Informatics. IEEE (2017). 978-1-5386-1107-4
Roy, S., Das, N., Kundu, M., Nasipuri, M.: Handwritten isolated Bangla compound character recognition: a new benchmark using a novel deep learning approach. In: Pattern Recognition Letters, vol. 90, pp. 15–21. Elsevier (2017)
Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional neural networks. In: Advances in Neural Information Systems, pp. 1097–1105 (2012)
Chandrakala, H.T., Thippeswamy, G., Gowda, S.D.: Enhancement of degraded document images using retinex and morphological operations. Int. J. Comput. Sci. Inf. Secur. 16(4) (2018)
Murty, R., Murthy, J.V.R., Prasad Reddy, P.V.G.D.: Text document classification based on a least square support vector machines with singular value decomposition. Int. J. Comput. Appl. (IJCA) 27(7), 21–26 (2011)
Acknowledgements
We thank the officials of Archaeological Survey of India (ASI), Mysore for permitting us to use the eleventh-century estampages of historical Kannada stone inscriptions from their archives for our research.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Chandrakala, H.T., Thippeswamy, G. (2020). Deep Convolutional Neural Networks for Recognition of Historical Handwritten Kannada Characters. In: Satapathy, S., Bhateja, V., Nguyen, B., Nguyen, N., Le, DN. (eds) Frontiers in Intelligent Computing: Theory and Applications. Advances in Intelligent Systems and Computing, vol 1014. Springer, Singapore. https://doi.org/10.1007/978-981-13-9920-6_7
Download citation
DOI: https://doi.org/10.1007/978-981-13-9920-6_7
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-13-9919-0
Online ISBN: 978-981-13-9920-6
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)