Slant Estimation and Correction for Online Handwritten Bengali Words
DOI:
https://doi.org/10.26438/ijcse/v6i5.535539Keywords:
Online Handwriting, Head Line, Base Line, Core Region, Virtical Lines, SlantAbstract
Slant is a common artefact inwhich handwritten word takes the form of slope and handwriting recognition system becomes less accurate. For this reason slant estimation and correction is a standard step in handwriting recognition systems for processing written text after skew detection and correction. If Skew correction and Slant correction are done successfully then recognition part will be more prominent. Handwritten characters of natural handwriting are usually italicized due to mechanism of handwriting and personality. In this paper the slant of the Bengali online handwritten words has been estimated and has been corrected. In slant estimation and correction, we have used Projection profile histogram method to detect core region or busy zone of the handwritten words and we defined the estimated head line (matra) and estimated base line of the words. Then we have detected almost vertical (considering a threshold value of angle 45 degree) straight lines which meet or closer to meet the head line (matra) and baseline of the core region present inside the words. After detecting those vertical straight lines we have calculated angle of slant of each vertical straight lines separately and calculated average slant angle. Then rotate all the pixels into the particular slant angle to do slant correction considering the base as fixed point. We have tested this proposed technique on 2655 Bengali handwritten words and have achieved an outstanding result of 96 percentage of accuracy.
References
Mazumdar, Bijaychandra,“The history of the Bengali language” (Repr. [d. Ausg.] Calcutta, 1920. ed.). New Delhi: Asian Educational Services. p. 57. ISBN 8120614526, 2000.
Fitrianingsih, Sarifuddin Madenda, Suryarini Widodo, Rodiah, “Slant Correction and Detection for Offline Cursive Handwriting using 2D Affine Transform”, International Journal of Engineering Research & Technology (IJERT), Vol. 5 Issue 08, August-2016.
A. Papandreou, B. Gatos, “Slant estimation and core-region detection for handwritten Latin words”, Pattern Recognition Letters, Vol. 35 (2014) , pp- 16–22.
Ch. N. Manisha, Y.K. Sundara Krishna, E. Sreenivasa Reddy, “Slant Correction for Offline Handwritten Telugu Isolated Characters and Cursive Words”, International Journal of Applied Engineering Research, Volume 11, Number 4 (2016) pp 2755-2760.
Alceu de S. Britto Jr, Robert Sabourin, Edouard Lethelier, Flávio Bortolozzi and Ching Y. Suen. (1999). “Slant normalization of handwritten numeral strings”. [Online] Available: www.etsmtl.ca/ETS/media/ImagesETS/Labo/LIVIA/Publications/1999/BrittoISKDM.pdf.
Ali, M. A., and Jumari, K. B., “Base-Area Detection and Slant Correction Techniques Applied for Arabic Handwritten Characters Recognition Systems”, Int. Conf. on Artificial Intelligence and Pattern Recognition, Orlando, USA. 2009. pp. 133-138.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors contributing to this journal agree to publish their articles under the Creative Commons Attribution 4.0 International License, allowing third parties to share their work (copy, distribute, transmit) and to adapt it, under the condition that the authors are given credit and that in the event of reuse or distribution, the terms of this license are made clear.
