Advantages of the Extended Water Flow Algorithm for Handwritten Text Segmentation

  • Darko Brodić
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6744)

Abstract

This paper identifies the advantages of the specific approach to water flow algorithm for multi-skewed handwritten text line segmentation. Original water flow algorithm assumes that hypothetical water flows, from both left and right sides of the document image frame, face obstruction from part of character, character, and group of characters in text lines. The stripes of areas left unwetted on the document image frame are finally labeled for the extraction of text lines. However, the method defines parameter water flow angle for flooding which depends on the text line slopes of each specific document. The estimation of the appropriate parameter value is difficult and limited as well. The limitation is manifested by possible election of only 4 values for this parameter. Extended approach has introduced enlargement of the parameter range. Consequently, decision making and the selection of the small values of the parameter below the minimum given by the original method shows improvement in the handwritten text line segmentation process. It is confirmed by the measurement on different types of letters.

Keywords

Image processing Document image processing Text line segmentation Handwritten text Water flow algorithm 

References

  1. 1.
    Likforman-Sulem, L., Zahour, A., Taconet, B.: Text Line Segmentation of Historical Documents: A Survey. IJDAR 9(2-4), 123–138 (2007)CrossRefGoogle Scholar
  2. 2.
    Amin, A., Wu, S.: Robust Skew Detection in Mixed Text/Graphics Documents. In: 8th ICDAR, Seoul, Korea, pp. 247–251 (2005)Google Scholar
  3. 3.
    Shi, Z., Govindaraju, V.: Line Separation for Complex Document Images using Fuzzy Runlength. In: Int. Workshop on Document Image Analysis for Libraries, Palo Alto, USA, pp. 306–312 (2004)Google Scholar
  4. 4.
    Basu, S., et al.: Text Line Extraction from Multi-skewed Handwritten Documents. Pattern Recognition 40(6), 1825–1839 (2007)CrossRefMATHGoogle Scholar
  5. 5.
    Brodic, D., Milivojevic, Z.: An Approach to Modification of Water Flow Algorithm for Segmentation and Text Parameters Extraction. In: Camarinha-Matos, L.M., Pereira, P., Ribeiro, L. (eds.) IFIP AICT, vol. 314, pp. 323–331. Springer, Heidelberg (2010)Google Scholar
  6. 6.
    Brodic, D., Milivojevic, Z.: Text Line Segmentation by Adapted Water Flow Algorithm. In: 10th Symposium NEUREL, Belgrade, Serbia, pp. 225–229 (2010)Google Scholar
  7. 7.
    Gonzalez, R.C., Woods, R.E.: Digital Image Processing, 2nd edn. Prentice-Hall, Englewood Cliffs (2002)Google Scholar
  8. 8.
    Otsu, N.: A Threshold Selection Method from Gray-level Histograms. IEEE Tran. on Systems, Man, and Cybernetics 9(1), 62–66 (1979)CrossRefGoogle Scholar
  9. 9.
    Sauvola, L., Pietikainen, M.: Adaptive Document Image Binarization. Pattern Recognition 33(2), 225–236 (2000)CrossRefGoogle Scholar
  10. 10.
    Bukhari, S.S., Shafait, F., Bruesl, T.M.: Adaptive Binarization of Unconstrained Hand-held Camera-captured Document Images. J. of Universal Computer Science 15(18), 3343–3363 (2009)Google Scholar
  11. 11.
    Khashman, A., Sekeroglu, B.: Document Image Binarisation Using a Supervised Neural Network. Int. J. of Neural Systems 18(5), 405–418 (2008)CrossRefGoogle Scholar
  12. 12.
    Preparata, F.P., Shamos, M.I.: Computational Geometry: An Introduction. Springer, Berlin (1985)CrossRefMATHGoogle Scholar
  13. 13.
    Chang, C.M.: Detecting Ellipses via Bounding Boxes. Asian Journal of Health and Information Sciences 1(1), 73–84 (2006)Google Scholar
  14. 14.
    Wang, J., Leung, M.K.H., Hui, S.C.: Cursive Word Reference Line Detection. Pattern Recognition 30(3), 503–511 (1997)CrossRefGoogle Scholar
  15. 15.
    Sanchez, A., et al.: Text Line Segmentation in Images of Handwritten Historical Documents. In: 1st IPTA, Sousse, Brasil, pp. 1–6 (2008)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Darko Brodić
    • 1
  1. 1.Technical Faculty BorUniversity of BelgradeBorSerbia

Personalised recommendations