HOME JOURNALS CONTACT

Asian Journal of Information Technology

Handwritten Bangla Numerical Digit Recognition using Histogram Technique
Md. Rezaul Bashar , Mirza A.F.M. Rashidul Hasan , Md. Altab Hossain and Dipankar Das

Abstract: This paper presents the Bangla Handwritten Digit Recognition system. In this developed system, the scanned image file of handwritten digit is taken as input. Image is then preprocessed. The boundary region of the image is detected and generates MxN vector to store all the color values of that image. Therefore, windowing technique is developed to extract feature of 16x16 from MxN matrix to produce a uniform feature. From the feature, histogram is produced and stored to a reference database. For recognition, the histogram of the recognizing digit is obtained in the same manner and a recognition engine compares with this recognizing histogram to the histograms of reference database and determines whether this system recognizes a digit or not. In this research, we have got a notable recognition rate.

How to cite this article
Md. Rezaul Bashar , Mirza A.F.M. Rashidul Hasan , Md. Altab Hossain and Dipankar Das , 2004. Handwritten Bangla Numerical Digit Recognition using Histogram Technique . Asian Journal of Information Technology, 3: 611-615.

© Medwell Journals. All Rights Reserved