Proportional Analysis of Feature Extraction for Tamil Handwritten Characters Using Centroid Based Method

A system for recognizing offline handwritten Tamil characters, using Feed Forward neural network, the primary objective of OHR is enhancement of a software for processing the images with different scripts and writing styles with high accuracy. The quality of the image is enhanced using diverse pre-processing operations; the feature of the character image is extracted using Centroid Based Distance Measure. In this paper a general description of the Algorithm is presented and classified using the Neural Network classifier. The proposed method is used for testing of handwritten Tamil character and has made a comparative analysis of different Distance Measure like Euclidean distance, Manhattan distance and Chessboard distance.