Asee peer logo

Segmentation Technique For The Learning In Recognition Of The Two Handwritten Bangla Digits Using Counterpropagation Neural Network

Download Paper |


2005 Annual Conference


Portland, Oregon

Publication Date

June 12, 2005

Start Date

June 12, 2005

End Date

June 15, 2005



Conference Session

New Approaches & Techniques in Engineering II

Page Count


Page Numbers

10.1100.1 - 10.1100.7



Permanent URL

Download Count


Request a correction

Paper Authors

author page

Abul Haque

author page

Afsaneh Minaie

Download Paper |

NOTE: The first page of text has been automatically extracted and included below in lieu of an abstract

Segmentation Technique for the Learning in Recognition of the Two Handwritten Bangla Digits Using Counterpropagation Neural Network

Abul L. Haque, Mohammed T. Siddique, Tanvir M. Khan, Imtiaz Ahmed Dept. of Computer Science & Engineering, North South University Dhaka, Bangladesh,,, Afsaneh Minaie, Engineering Department, Utah Valley State College,

Abstract We are proposing a segmentation technique for the learning of the two Bangla digits incorporating the grid method, the regional search method, the feature extraction method, and the counterpropagation neural network. We emphasized the recognition of the two Bangla digits one and nine since these two digits look similar. The experimental result shows an overall increased rate of recognition for Bangla one and nine.

Keywords: Bangla numeral, Bangla digit, Neural Network, Counterpropagation Neural Network, segmentation

1. Introduction Bangla is a language spoken by the people of Bangladesh and the people of the state of West Bengal in India. Bangla character recognition using different approaches has become a field of interest in the last few years. In spite of its interests, the progress in this field has not been very fast. So far, only a limited amount of work has been done successfully with Bangle character recognition due to its complexity. Literature survey on the Bangla digit recognition techniques revealed that the simulation results of all the approaches were suffering from the low recognition rates of Bangla one and nine since both of them look similar. Therefore, we paid special attention on the recognition of Bangla digit one and nine. Earlier work of Rahman et al.9, Syeed et al.10, Mollah and Talukder11, and Rahman et al.12 inspired us to initiate the work in this area. The technique that was formulated here for the improvement of the recognition incorporated the

Proceedings of the 2005 American Society for Engineering Education Annual Conference & Exposition t Copyright © 2005, American Society for Engineering Education

Haque, A., & Minaie, A. (2005, June), Segmentation Technique For The Learning In Recognition Of The Two Handwritten Bangla Digits Using Counterpropagation Neural Network Paper presented at 2005 Annual Conference, Portland, Oregon. 10.18260/1-2--14338

ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2005 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015