Performance Analysis on Bangla and Arabic Numeral Recognition

Gita Sinha, Md. Ashif Habibi

Abstract: Handwritten numeral recognition is a challenging work at this time. In this paper we deal with accuracy comparison among Arabic numerals and Bangla numeral recognition. This paper composed with four main phases pre-processing, feature extraction, classification and result comparison with both numerals. After pre-processing image are divided into different equal parts and then applied feature extraction techniques. Numerical feature extracted contain more details about image that provide improved recognition accuracy. These numerals recognition is carried out by SVM classifier.

Keywords: Handwritten and printed numeral recognition, features extraction, pre-processing, support vector machine, image centroid zone and zone centroid zone, handwritten numeral recognition

How to Cite?: Gita Sinha, Md. Ashif Habibi, "Performance Analysis on Bangla and Arabic Numeral Recognition", Volume 6 Issue 1, January 2017, International Journal of Science and Research (IJSR), Pages: 1536-1539, https://www.ijsr.net/getabstract.php?paperid=ART20164370, DOI: https://dx.doi.org/10.21275/ART20164370