Handwritten Hindi character recognition using multilayer perceptron and radial basis function neural networks

This paper compares the multilayer perceptron (MLP) networks and the radial basis function (RBF) networks in the task of handwritten Hindi character recognition (HCR). The error backpropagation algorithm was used to train the MLP networks. An automatic HCR system using MLP and RBF networks is presen...

Full description

Saved in:
Bibliographic Details
Main Author: Verma, B.K.
Format: Conference Proceeding
Language:English
Subjects:
Backpropagation algorithms
Character recognition
Feature extraction
Multi-layer neural network
Multilayer perceptrons
Natural languages
Neural networks
Pattern recognition
Pixel
Radial basis function networks
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper compares the multilayer perceptron (MLP) networks and the radial basis function (RBF) networks in the task of handwritten Hindi character recognition (HCR). The error backpropagation algorithm was used to train the MLP networks. An automatic HCR system using MLP and RBF networks is presented. The experiments were carried out on two hundred forty five samples of five writers. The results showed that the MLP networks trained by the error backpropagation algorithm were superior in recognition accuracy and memory usage. However, they suffered from long training time than that of RBF networks.
DOI:10.1109/ICNN.1995.489003