A SVM Greek character recogniser. (11th May 2009)

• Record Type: Journal Article
• Authors: Camastra, Francesco
• Is Part Of: International journal of intelligent defence support systems. Volume 1:Number 4(2008)
• Page Start: 290
• Page End: 299
• Publication Date: 2009-05-11
• Subjects: SVMs -- support vector machines -- LVQ -- learning vector quantisation -- MLP -- multi-layer perceptron -- Greek character recognition -- feature extraction -- classifiers; Artificial intelligence -- Military applications -- Periodicals; Cybernetics -- Periodicals; Experimental design -- Periodicals; 623.028563
• Languages: English
• ISSNs: 1755-1587

Abstract

This paper presents a handwritten Greek character recogniser based on Support Vector Machines (SVMs). The recogniser is composed of two modules: the first one is a feature extractor, the second one, the classifier, is performed by means of SVMs. The recogniser, tested on a database of more than 22000 handwritten Greek characters, has shown satisfactory performances. SVMs compare notably better, in terms of recognition rates, with popular neural classifiers, such as Learning Vector Quantisation (LVQ) and Multi-layer Perceptron (MLP).