The Effect of Pre-processing and Testing Methods on Online Kannada Handwriting Recognition: Studies Using Signal Processing and Statistical Techniques

Ramya, S and Kumara, Shama (2018) The Effect of Pre-processing and Testing Methods on Online Kannada Handwriting Recognition: Studies Using Signal Processing and Statistical Techniques. Pertanika Journal of Science and Technology, 26 (2). pp. 671-690. ISSN 01287680

[img] PDF
4383.pdf - Published Version
Restricted to Registered users only

Download (2MB) | Request a copy
Official URL: http://www.pertanika.upm.edu.my/

Abstract

Pre-processing and testing methodology plays a significant role in online handwritten character recognition. Although many researchers have proposed several pre-processing and testing methods, the effect of these techniques on the recognition and comparisons among them are ignored. In this work, experiments were conducted to analyse the effect of various pre-processing and testing methods on Kannada handwritten data. The focus of the present work is to statistically quantify the effect on recognition time and accuracy through experiments using different pre-processing methods on online handwritten data processed by the Support Vector Machine (SVM). The performance of the SVM is also compared with various other training and testing methodology. The performance of the online handwriting recognition system is affected dramatically by the various pre-processing and testing methods. Stratified tenfold cross validation showed better performance for the Kannada handwritten dataset

Item Type: Article
Uncontrolled Keywords: Bootstrapping, cross-validation, down sampling, normalisation, online handwriting recognition, resampling, single partition testing, smoothing
Subjects: Engineering > MIT Manipal > Electronics and Communication
Depositing User: MIT Library
Date Deposited: 22 Jun 2018 04:46
Last Modified: 22 Jun 2018 04:46
URI: http://eprints.manipal.edu/id/eprint/150945

Actions (login required)

View Item View Item