Abstract
In this paper, we propose a word segmentation method that is based on fringe maps on Telugu script. Our objective is to create a data set of word images for enabling direct training for recognition on those. The standard methods employed for the task of word segmentation in Telugu OCR systems are projection profiles and run-length smearing. However those methods have their limitations. In this work a different application of fringe maps is shown for line segmentation into words. Fringes were previously applied successfully for carrying out classification and line segmentation. Telugu script, which has consonant modifiers that are usually placed below or below-right to the base consonants. This kind of orthographic property leads to characters that may touch each other. One way to deal with touched characters is to make use of segmentation free methods, which do not need prior segmentation of word images into characters or connected components. The novelty of our method is that we analyze fringe maps of document images to find an appropriate fringe value threshold and apply it for word segmentation of Telugu documents. Encouraging results are observed with our fringe value threshold based word segmentation. We observe that choosing higher threshold fringe values leads to under-segmentation of words, whereas lower values cause over-segmentation of words. Our word segmentation approach is successfully compared with the widely used projection profiles based word segmentation method.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Achanta, R., Hastie, T.: Telugu OCR framework using deep learning. CoRR abs/1509.05962 (2015)
Bhagvati, C., Ravi, T., Kumar, S.M., Negi, A.: On developing high accuracy OCR systems for Telugu and other Indian scripts. In: 2002 Proceedings of the Language Engineering Conference, pp. 18–23, December 2002
Brown, R.L.: The fringe distance measure: an easily calculated image distance measure with recognition results comparable to Gaussian blurring. IEEE Trans. Syst. Man Cybern. 24(1), 111–115 (1994)
Koppula, V.K., Negi, A.: Fringe map based text line segmentation of printed Telugu document images. In: 2011 International Conference on Document Analysis and Recognition, pp. 1294–1298, September 2011
Krishnan, P., Sankaran, N., Singh, A.K., Jawahar, C.V.: Towards a robust OCR system for Indic scripts. In: 2014 11th IAPR International Workshop on Document Analysis Systems (DAS), pp. 141–145, April 2014
Kumar, P.P., Bhagvati, C., Negi, A., Agarwal, A., Deekshatulu, B.L.: Towards improving the accuracy of Telugu OCR systems. In: ICDAR, pp. 910–914. IEEE Computer Society (2011)
Lakshmi, C.V., Patvardhan, C.: An optical character recognition system for printed Telugu text. Pattern Anal. Appl. 7(2), 190–204 (2004)
Negi, A., Bhagvati, C., Krishna, B.: An OCR system for Telugu. In: ICDAR, pp. 1110–1114. IEEE Computer Society (2001)
Rao, D.K., Negi, A.: Orthographic properties based Telugu text recognition using hidden Markov models. In: 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), vol. 05, pp. 45–50, November 2017
Vasantha Lakshmi, C., Patvardhan, C.: A multi-font OCR system for printed Telugu text. In: 2002 Proceedings of the Language Engineering Conference, pp. 7–17, December 2002
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Devarapalli, K.R., Negi, A. (2019). Telugu Word Segmentation Using Fringe Maps. In: Sundaram, S., Harit, G. (eds) Document Analysis and Recognition. DAR 2018. Communications in Computer and Information Science, vol 1020. Springer, Singapore. https://doi.org/10.1007/978-981-13-9361-7_8
Download citation
DOI: https://doi.org/10.1007/978-981-13-9361-7_8
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-13-9360-0
Online ISBN: 978-981-13-9361-7
eBook Packages: Computer ScienceComputer Science (R0)