Recognition of Handwritten Bangla Numerals using Adaptive Coefficient Matching Technique

Abstract

This paper presents an approach to recognize handwritten bangla numerals from low pixel mobile camera based on correlation co-efficient. In this paper, an approach is identified to recognize such handwritten bangla numerals having different shape and scan through from any low pixel mobile device. It has also shown that while writing in continues sometimes two separated digit connected to each other, to distinguish such digit we used Line segmentation. High correlation coefficient will provide the successful results of recognition. A large dataset is used for training and testing purpose which gives overall 93.80% accuracy using mean of maximum correlation and 95.70% accuracy is shown when individually correlation found.