A Genetic Algorithm for Feature Selection in a Neuro-Fuzzy OCR System

Authors:
Affiliations:
Venue:
ICDAR '01 Proceedings of the Sixth International Conference on Document Analysis and Recognition
Year:
2001

Citing 0
Cited 4

Exploiting Reliability for Dynamic Selection of Classifiers by Means of Genetic Algorithms

ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 2
Novel Framework for Selecting the Optimal Feature Vector from Large Feature Spaces

ICIAR '09 Proceedings of the 6th International Conference on Image Analysis and Recognition
Parametric architecture for function calculation improvement

ARCS'07 Proceedings of the 20th international conference on Architecture of computing systems
A genetic algorithm based region sampling for selection of local features in handwritten digit recognition application

Applied Soft Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Abstract: We have worked on the development of a character recognition system in the soft computing paradigm. In this paper we present a genetic algorithm used for feature selection with a Feature Quality Index (FQI) metric. We generate feature vectors by defining fuzzy sets on Hough transform of character pattern pixels. Each feature element is multiplied by a mask vector bit before reaching the input of a multilayer perceptron (MLP). The genetic algorithm operates on the bit string represented by the mask vector to select the best set of features. The method has been tested with three benchmark data sets and the results show a fast convergence of the genetic algorithm.