Volume : IV, Issue : IV, April - 2014

Pre–processing and Self training techniques in Handwritten Character Recognition

Magesh Kasthuri, Dr. V. Shanthi

Abstract :

 Handwriting recognition is a widely implemented technology to electronically identify handwritten text. In online handwriting recognition, text written on a touch surface of an electronic touch device is dynamically recognized based on the movements of a writing device (digital pen, a finger, or a stylus) and presented to the user after each continuous stroke such as a character or word is entered.

Therefore, the user is able to edit a character or a word in the event of incorrect recognition of text by the electronic device as soon as it is presented to the user. On the other hand, in offline handwriting recognition, recognition of characters occurs by identifying characters from an image of handwritten text instead of the user writing on a touch surface. Here, the recognized text is presented to a user only when the entire handwritten text represented in the image is recognized.

An Artificial Intelligence based Neural system is successful when it is trained properly. The basis of training a neural network is to prepare the system to recognize various combinations of possibilities of characters. From a larger picture this is the key requirement for training a neural network. But there are underlying principles in training a neural network.

In other words, training is not a simple record and process technique but also a technique which decides the working logic of the processing algorithm.

This paper discusses some of the key features of training a neural network along with some key training methodologies for Handwritten character recognition system.

Keywords :

Article: Download PDF   DOI : 10.36106/ijar  

Cite This Article:

"Magesh Kasthuri, Dr.V.Shanthi Pre - processing and Self training techniques in Handwritten Character Recognition Indian Journal of Applied Research, Vol.IV, Issue. IV

Number of Downloads : 934

References :