International Journal of Computer Science & Network Security

International Journal of Computer Science & Network Security (국제컴퓨터통신보호논문지학회)
권호 표지
DOI QR코드

DOI QR Code

Handwritten Indic Digit Recognition using Deep Hybrid Capsule Network

  • Received : 2024.02.05
  • Accepted : 2024.02.20
  • Published : 2024.02.29
Download PDF
⟨ Previous Next ⟩

Abstract

Indian subcontinent is a birthplace of multilingual people where documents such as job application form, passport, number plate identification, and so forth is composed of text contents written in different languages/scripts. These scripts may be in the form of different indic numerals in a single document page. Due to this reason, building a generic recognizer that is capable of recognizing handwritten indic digits written by diverse writers is needed. Also, a lot of work has been done for various non-Indic numerals particularly, in case of Roman, but, in case of Indic digits, the research is limited. Moreover, most of the research focuses with only on MNIST datasets or with only single datasets, either because of time restraints or because the model is tailored to a specific task. In this work, a hybrid model is proposed to recognize all available indic handwritten digit images using the existing benchmark datasets. The proposed method bridges the automatically learnt features of Capsule Network with hand crafted Bag of Feature (BoF) extraction method. Along the way, we analyze (1) the successes (2) explore whether this method will perform well on more difficult conditions i.e. noise, color, affine transformations, intra-class variation, natural scenes. Experimental results show that the hybrid method gives better accuracy in comparison with Capsule Network.

Keywords

References

  1. A. Roy, "Indian shield: Pristine shape, size and tectonic framework," in Geological Evolution of the Precambrian Indian Shield, pp. 1-15, Springer, 2019.
  2. B. E. Sawe, What Language Is Spoken in India? - WorldAtlas.com, July 10, 2018 (Accessed July 3, 2019). https://www.worldatlas.com/articles/the-most-widely-spoken-languages-in-india.html.
  3. Indian Languages-Defining India's Internet - KPMG International Cooperative [NL], April 25, 2017 (Accessed July 4, 2019). https://assets.kpmg/content/dam/kpmg/in/pdf/2017/04/Indian-languages-Defining-Indias-Internet.pdf.
  4. S. Pratt, A. Ochoa, M. Yadav, A. Sheta, and M. Eldefrawy, "Handwritten digits recognition using convolution neural networks," The Journal of Computing Sciences in Colleges, p. 40, 2019.
  5. B. Lopez, M. A. Nguyen, and A. Walia, "Modified mnist," 2019.
  6. S. Majumder, C. von der Malsburg, A. Richhariya, and S. Bhanot, "Handwritten digit recognition by elastic matching," arXiv preprint arXiv:1807.09324, 2018.
  7. B. N. Dhannoon and H. H. Al, "Handwritten hindi numerals recognition," International Journal of Innovation and Applied Studies, 05 2013.
  8. M. Chaudhary, M. H. Mirja, and N. Mittal, "Hindi numeral recognition using neural network," Int. J. Sci. Eng. Res., vol. 5, no. 6, pp. 260-268, 2014.
  9. G. Singh and S. Lehri, "Recognition of handwritten hindi characters us- ing backpropagation neural network," International Journal of Computer Science and Information Technologies, vol. 3, no. 4, pp. 4892-4895, 2012.
  10. R. Noor, K. M. Islam, and M. J. Rahimi, "Handwritten bangla numeral recognition using ensembling of convolutional neural network," in 2018 21st International Conference of Computer and Information Technology (ICCIT), pp. 1-6, IEEE, 2018.
  11. M. Kumar, M. Jindal, R. Sharma, and S. R. Jindal, "Performance eval- uation of classifiers for the recognition of offline handwritten gurmukhi characters and numerals: a study," Artificial Intelligence Review, pp. 1- 23, 2019.
  12. Pauly, R. D. Raj, and B. Paul, "Hand written digit recognition system for south indian languages using artificial neural networks," in 2015 Eighth International Conference on Contemporary Computing (IC3), pp. 122-126, IEEE, 2015.
  13. J. M. Alghazo, G. Latif, L. Alzubaidi, and A. Elhassan, "Multi-language handwritten digits recognition based on novel structural features," Journal of Imaging Science and Technology, vol. 63, no. 2, pp. 20502-1, 2019
  14. V. U. Prabhu, S. Han, D. A. Yap, M. Douhaniaris, P. Seshadri, and J. Whaley, "Fonts-2-handwriting: A seed-augment-train framework for universal digit classification," arXiv preprint arXiv:1905.08633, 2019.
  15. M. Z. Alom, P. Sidike, T. M. Taha, and V. K. Asari, "Handwritten bangla digit recognition using deep learning," arXiv preprint arXiv:1705.02680, 2017.
  16. M. Z. Alom, P. Sidike, T. M. Taha, and V. K. Asari, "Handwritten bangla digit recognition using deep learning," arXiv preprint arXiv:1705.02680, 2017.
  17. S. Sabour, N. Frosst, and G. E. Hinton, "Dynamic routing between cap- sules," in Advances in neural information processing systems, pp. 3856- 3866, 2017.
  18. Y. LeCun, C. Cortes, and C. Burges, "Mnist handwritten digit database. 2010," URL http://yann.lecun.com/exdb/mnist, vol. 3, no. 1, 2010.
  19. S. O'Hara and B. A. Draper, "Introduction to the bag of fea- tures paradigm for image classification and retrieval," arXiv preprint arXiv:1101.3354, 2011.
  20. E. Mayoraz and E. Alpaydin, "Support vector machines for multi-class classification," in International Work-Conference on Artificial Neural Networks, pp. 833-842, Springer, 1999.
  21. J. P. Jones and L. A. Palmer, "An evaluation of the two-dimensional gabor filter model of simple receptive fields in cat striate cortex," Journal of neurophysiology, vol. 58, no. 6, pp. 1233-1258, 1987. https://doi.org/10.1152/jn.1987.58.6.1233
  22. C. Cortes and V. Vapnik, "Support-vector networks," Machine learning, vol. 20, no. 3, pp. 273-297, 1995.
  23. LeCun et al., "Lenet-5, convolutional neural networks," URL: http://yann. lecun.com/exdb/lenet, vol. 20, p. 5, 2015.
  24. G. E. Hinton, S. Osindero, and Y.-W. Teh, "A fast learning algorithm for deep belief nets," Neural computation, vol. 18, no. 7, pp. 1527-1554, 2006. https://doi.org/10.1162/neco.2006.18.7.1527
  25. Yosinski, J. Clune, Y. Bengio, and H. Lipson, "How transferable are features in deep neural networks?," in Advances in neural information processing systems, pp. 3320- 3328, 2014
  26. K. Nogueira, O. A. Penatti, and J. A. dos Santos, "Towards better exploiting convolutional neural networks for remote sensing scene classification," Pattern Recognition, vol. 61, pp. 539-556, 2017. https://doi.org/10.1016/j.patcog.2016.07.001
  27. N. Das, J. M. Reddy, R. Sarkar, S. Basu, M. Kundu, M. Nasipuri, and D. K. Basu, "A statistical-topological feature combination for recognition of handwritten numerals," Applied Soft Computing, vol. 12, no. 8, pp. 2486-2495, 2012
  28. N. Das, R. Sarkar, S. Basu, M. Kundu, M. Nasipuri, and D. K. Basu, "A genetic algorithm based region sampling for selection of local features in handwritten digit recognition application," Applied Soft Computing, vol. 12, no. 5, pp. 1592-1606, 2012. https://doi.org/10.1016/j.asoc.2011.11.030