Hand-written English numeral recognition system using neural network

Abstract

This thesis aims at implementing an algorithm for recognition of hand-written English numeral. Handwriting recognition has been one of the active and challenging research areas in the field of image processing and pattern recognition. In this thesis the digits are classified into two groups, one group comprises of blobs with/without stems and the other digits with stems only. The blobs are identified based on a new concept called morphological region filling technique. This eliminates the issue of finding the size of blobs and their structuring elements. This method completely eliminates the complex process of recognition of horizontal or vertical lines. This extracted feature will then classified with the help of neural network train tool. It is a faster English numeral recognition algorithm it uses part of the character instead of complete image.