Research Article Open Access

Performance Comparison of Different Image Sizes for Recognizing Unconstrained Handwritten Tamil Characters using SVM

N. Shanthi and K. Duraiswamy

Abstract

This study describes a system for recognizing offline handwritten Tamil characters using Support Vector Machine (SVM). Data samples are collected from different writers on A4 sized documents. They are scanned using a flat bed scanner at a resolution of 300 dpi and stored as grey scale images. Various preprocessing operations are performed on the digitized image to enhance the quality of the image. Random sized preprocessed image is normalized to uniform sized image. Pixel densities are calculated for different zones of the image and these values are used as the features of a character. These features are used to train and test the support vector machine. The support vector machine is tested for the first time for recognizing handwritten Tamil characters. The recognition results are tested for 3 different standard sizes of 32X32, 48X48 and 64X64. Pixel densities are calculated for various zones and also for overlapping zones of the 64X64 sized image. Best results are obtained for 64X64 sized normalized image with overlapping windows. The handwriting system is trained for 106 different characters and test results are given for 34 different Tamil characters. With a simple feature of pixel density, the system has achieved a very good recognition rate of 87.4% on the totally unconstrained handwritten Tamil character database.

Journal of Computer Science
Volume 3 No. 9, 2007, 760-764

DOI: https://doi.org/10.3844/jcssp.2007.760.764

Submitted On: 16 August 2007 Published On: 30 September 2007

How to Cite: Shanthi, N. & Duraiswamy, K. (2007). Performance Comparison of Different Image Sizes for Recognizing Unconstrained Handwritten Tamil Characters using SVM. Journal of Computer Science, 3(9), 760-764. https://doi.org/10.3844/jcssp.2007.760.764

  • 2,553 Views
  • 2,262 Downloads
  • 15 Citations

Download

Keywords

  • Tamil character recognition
  • support vector machine
  • preprocessing
  • feature extraction