Academic Journals Database
Disseminating quality controlled scientific knowledge

A Comparative Study using Wavelet and SVM for Devanagri Characters Recognition

ADD TO MY LIST
 
Author(s): Anil Kumar N. Holambe | Ravinder C. Thool | Sushil Kumar N.Holambe | Ganesh K. Pakle

Journal: International Journal of Advanced Research in Computer Science
ISSN 0976-5697

Volume: 01;
Issue: 04;
Start page: 69;
Date: 2010;
Original page

Keywords: Wavelet | MAT | Complex | Wavelet | Autocorrelation.

ABSTRACT
This paper presents a wavelet-based approach for recognizing handwritten and printed Devnagari characters. In this paper we have used wavelet for feature extraction of the character. We have developed six handwritten data feature sets and six printed data feature set ,each dataset is divided in four part for our experimentation. We have also used wavelet kernels and regular kernels in SVM classification. Each SVM kernel is applied on total 12 x4=48 feature datasets.
Affiliate Program      Why do you need a reservation system?