Tamil Character Recognition

Abstract

Ancient Tamil character recognition is an important research and application
area on pattern recognition theory, which plays an important role on realizing document.
The main objective of the project is to determine characters from any given text of A-Z and
Tamil consonants. A system of English and Ancient Eighth Century Tamil character
recognition based on point feature extraction of the character. Eighth Century Tamil
characters are recognized using Multilayer Feed Forward Neural Network has been
described in our work. Pre-processing is the first step which is done using thresholding
algorithm. Feature extraction forms the basic underlying part of recognizing each character
during testing. Characters are classified by Back Propagation Neural Network. An Artificial
Neural Network has been created and trained to diagnose a single character. A Feature
Vector of each character has been used to train the Neural Network. This system will be
suitable for converting handwritten documents into structural text format. The
handwritten character recognition system with better-quality recognition rates will be
eminently suitable for several applications including document reading and conversion of
any handwritten document into structural text format.