Printed Gujarati Character Classification Using High-Level Strokes

  • Mukesh M. Goswami
  • Suman K. Mitra
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 704)


This paper presents a method of extracting and identifying the high-level strokes from the off-line printed Gujarati characters. The ability to provide incremental definition of characters as a sequence of high-level strokes makes the proposal unique. The features are validated on subset of printed Gujarati characters using probabilistic classifiers. The validation dataset consists of samples from three different sources, namely machine-printed books, newspapers, and laser-printed documents in order to ensure varieties of ink thickness, size, and font type and style. Classification is performed first using simple Naive Bayes classifier to test the discriminating strength of features. The results are further improved by taking first-order dependency between high-level strokes using hidden Markov models. The average recognition accuracy obtained on different datasets is in between 93 to 96% using hidden Markov model, which is comparable with other existing work. The experimental result shows that the high-level strokes are independent of font type and style as well as provides highly compact and sequential representation of high-level shape of characters which is also useful in other applications like shape-based word-image matching.


Stroke features Character classification Gujarati characters 


  1. 1.
    Albus, J., Anderson, R., Brayer, J., DeMori, R., Feng, H., Horowitz, S., Moayer, B., Pavlidis, T., Stallings, W., Swain, P., Vamos, T.: Syntactic pattern recognition, applications, vol. 14. Springer Science & Business Media (2012)Google Scholar
  2. 2.
    Antani, S., Agnihotri, L.: Gujarati character recognition. In: Proc. of the 5th Int. Conf. on Document Analysis and Recognition (ICDAR’99). pp. 418–421 (1999)Google Scholar
  3. 3.
    Bayes, T., Richard, P.: An essay towards solving a problem in the doctrine of chances. Philosophical Transactions of the Royal Society of London 53, 370–418 (1763)Google Scholar
  4. 4.
    Chaudhuri, B. (ed.): Digital Document Processing: Major Directions and Recent Advances. Springer-Verlag London Ltd (2007)Google Scholar
  5. 5.
    Dholakia, J., Negi, A., Mohan, S.: Progress in Gujarati document processing and character recognition. In: Govindaraju, V., Setlur, S. (eds.) Guide to OCR for Indic Scripts: Document Recognition and Retrieval, pp. 73–95. Springer Publishing Company (2009)Google Scholar
  6. 6.
    Dholakia, J., Yajnik, A., Negi, A.: Wavelet feature based confusion character sets for Gujarati script. In: Proc. of the Int. Conf. on Computational Intelligence and Multimedia Applications. pp. 366–370 (2007)Google Scholar
  7. 7.
    Garofalakis, M., Rastogi, R., Shim, K.: Spirit: Sequential pattern mining with regular expression constraints. VLDB 99, 7–10 (1999)Google Scholar
  8. 8.
    Goswami, M., Mitra, S.K.: Classification of printed Gujarati characters using low-level stroke features. ACM Trans. Asian Low-Resour. Lang. Inf. Process. 15(4), 25:1–26 (2015)Google Scholar
  9. 9.
    Goswami, M., Prajapati, H., Dabhi, V.: Classification of printed Gujarati characters using SOM based k-nearest neighbor classifier. In: Proc. of the Int. Conf. on Image Information Processing. pp. 1–5. IEEE (2011)Google Scholar
  10. 10.
    Govindaraju, V., Setlur, S. (eds.): Guide to OCR for Indic Scripts: Document Recognition and Retrieval. Springer Publishing Company (2009)Google Scholar
  11. 11.
    Hassan, E., Chaudhury, S., Gopal, M.: Feature combination for binary pattern classification. International Journal of Document Analysis and Recognition (IJDAR) 17(4), 375–392 (2014)Google Scholar
  12. 12.
    Jayadevan, R., Kolhe, S., Patil, P., Pal, U.: Off-line recognition of Devanagari script: A survey. IEEE Trans. Syst. Man Cybern., Part C (Applications and Reviews) 41(6), 782–796 (Nov 2011)Google Scholar
  13. 13.
    Jha, G.: The TDIL program and the Indian language corpora initiative (ILCI). In: Proc. of the 7th Conf. on Int. Language Resources and Evaluation (LREC 2010). European Language Resources Association (ELRA) (2010)Google Scholar
  14. 14.
    Koller, D., Friedman, N.: Probabilistic graphical models: principles and techniques. MIT Press (2009)Google Scholar
  15. 15.
    Kompalli, S., Setlur, S., Govindaraju, V.: Challenges in OCR of Devanagari documents. In: Proc. of the 8th Int. Conf. on Document Analysis and Recognition (ICDAR’05). pp. 1–5. IEEE (2005)Google Scholar
  16. 16.
    Mandar, C., Gitam, M., Suman, M., Mukesh, G.: Similar looking Gujarati printed character recognition using locality preserving projection and artificial neural network. In: Proc. of the 3rd Int. Conf. on Emerging Applications of Information Technology (EAIT’12). pp. 457–461. IEEE (2012)Google Scholar
  17. 17.
    Martin, J.C.: Introduction to Languages and the Theory of Computation. McGraw-Hill (1991)Google Scholar
  18. 18.
    Murphy, K.: Machine Learning: A Probabilistic Perspective. The MIT Press, Cambridge, Massachusetts London, England (2012)Google Scholar
  19. 19.
    Pal, U., Chaudhuri, B.: Indian script character recognition: a survey. Pattern Recognition 37(9), 1887–1899 (Sep 2004)Google Scholar
  20. 20.
    Pal, U., Ramachandran, J., Sharma, N.: Handwriting recognition in Indian regional scripts: a survey of off-line techniques. ACM Transactions on Asian Language Information Processing (TALIP) 11(1), A1–35 (2012)Google Scholar
  21. 21.
    Peeta, B., Ramakrishnan, A.: OCR in Indian scripts : A survey. IETE Technical Review 22(3), 217–227 (2005)Google Scholar
  22. 22.
    Rabiner, L.: A tutorial on Hidden Markov Models and selected applications in speech recognition. Proc. of the IEEE 77(2), 256–287 (1989)Google Scholar
  23. 23.
    Sharma, A., Shah, S.: Design and implementation of optical character recognition system to recognize Gujarati script using template matching. IE(I) Journal(ET) 86(1), 44–49 (2006)Google Scholar
  24. 24.
    Soumen, B., Gaurav, H.: A survey on optical character recognition for Bangla and Devanagari scripts. Sadhana 38(1), 133–168 (Feb 2013)Google Scholar
  25. 25.
    Welch, L.: Hidden Markov Models and the Baum-Welch algorithm. IEEE Information Theory Society Newsletter 53, 10–13 (2003)Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  1. 1.Department of IT, Faculty of TechnologyD. D. UniversityNadiadIndia
  2. 2.Dhirubhai Ambani Institute of Information and Communication TechnologyGandhinagarIndia

Personalised recommendations