Feature Extraction and Recognition of Ancient Kannada Epigraphs

Conference paper
Part of the Smart Innovation, Systems and Technologies book series (SIST, volume 33)


Optical Character Recognition finds numerous applications and one among them is in the field of Epigraphy, which is the study of inscriptions. Expert epigraphers who read ancient inscriptions are nowadays less in number. Also it is found that preserving these ancient records is important, hence lot of scope for digitization of these historical records and automatic decipherment of the same is important to the mankind. This paper addresses mainly on Segmentation, Feature Extraction and Character Recognition of Ancient Kannada script of Ashoka and Hoysala periods. Initially, input epigraph image is segmented to obtain sampled characters using Nearest Neighbor clustering Algorithm. Statistical Features such as Mean, Variance, Standard Deviation, Kurtosis, Skewness, Homogeneity, Contrast, Correlation, Energy, and Coarseness are extracted to store as training set and for comparison at the later stage of testing. Finally Mamdani Fuzzy Classifier is used in recognition; as a result, output is displayed in modern Kannada form.


Ancient Kannada script Nearest neighbor clustering Statistical features Optical character recognition Fuzzy classifier 


  1. 1.
    Bandara, D., Warnajith, N., Minato, A., Ozawal, S.: Creation of precise alphabet fonts of early Brahmi script from photographic data of ancient Sri Lankan inscriptions. Can. J. Artif. Intell. Mach. Learn. Pattern Recognit. 3(3), 33–39 (2012)Google Scholar
  2. 2.
    Kumar, S.R., Bharathi, V.S.: An off line ancient tamil script recognition from temple wall inscription using Fourier and Wavelet features. Eur. J. Sci. Res. 80(4), 457–464 (2012). ISSN 1450-216XGoogle Scholar
  3. 3.
    Soumya, A., Kumar, G.H.: Dating of ancient epigraphs using random forest classifier. In: International Conference on Emerging Computation and Information Technologies, ICECIT-2013, pp. 331–339. Elsevier publications, Siddaganga Institute of Technology, Tumkur (2013)Google Scholar
  4. 4.
    Soumya, A., Kumar, G.H.: Zernike moment features for the recognition of ancient Kannada base characters. Int. J. Graph. Image Process. (IJGIP) 4(2), 99–104 (2014)Google Scholar
  5. 5.
    Yadav, N., Joglekar, H., Rao, R.P.N., Vahia, M.N., Adhikari, R., Mahadevan, I.: Statistical analysis of the indus script using n-grams. PLoS One 5(3), 1–15 (2010)CrossRefGoogle Scholar
  6. 6.
    Yang, M.Q., Kidiyo, K., Joseph, R.: A survey of shape feature extraction techniques, Yin, P.-Y. (ed.). Pattern Recogn 12(1), 43–90 (2008)Google Scholar
  7. 7.
    Murthy, K.S., Kumar, G.H., Kumar, P.S., Ranganath, P.R.: Nearest neighbor clustering based approach for line and character segmentation in epigraphical scripts. In: International Conference on Cognitive Systems, New Delhi, 14–15 Dec 2004Google Scholar
  8. 8.
    Elamvazuthi, Vasant, P., Webb, J.: The application of Mamdani fuzzy model for auto zoom function of a digital camera. IJCSIS 6(3), 126–134 (2009)Google Scholar

Copyright information

© Springer India 2015

Authors and Affiliations

  1. 1.Department of Computer Science and EngineeringR V College of EngineeringBangaloreIndia
  2. 2.Department of Studies in Computer ScienceUniversity of MysoreMysoreIndia

Personalised recommendations