Handwritten Arabic (Indian) Numerals Recognition Using Fourier Descriptor and Structure Base Classifier

Abstract

In this paper a simple and accurate method is proposed to recognize Arabic (Indian) numerals using Fourier descriptors as the main classifier feature set, and to improve the recognition accuracy a simple structure based classifier is add as a supplementary classifier. The recognition system was tested on 450 samples collected from 5 students and the test results indicate that the recognition ratio is%89.6when only 5 Fourier descriptors are used as discriminating features set, and the ratio is raised to%98 when 4 Fourier descriptors are used in addition to the simple structure based classifier.