Ambiguity Reduction Through Optimal Set of Region Selection Using GA and BFO for Handwritten Bangla Character Recognition

Ambiguity Reduction Through Optimal Set of Region Selection Using GA and BFO for Handwritten Bangla Character Recognition

Nibaran Das (Jadavpur University, India), Subhadip Basu (Jadavpur University, India), Mahantapas Kundu (Jadavpur University, India) and Mita Nasipuri (Jadavpur University, India)
Copyright: © 2018 |Pages: 28
DOI: 10.4018/978-1-5225-5204-8.ch053

Abstract

To recognize different patterns, identification of local regions where the pattern classes differ significantly is an inherent ability of the human cognitive system. This inherent ability of human beings may be imitated in any pattern recognition system by incorporating the ability of locating the regions that contain the maximum discriminating information among the pattern classes. In this chapter, the concept of Genetic Algorithm (GA) and Bacterial Foraging Optimization (BFO) are discussed to identify those regions having maximum discriminating information. The discussion includes the evaluation of the methods on the sample images of handwritten Bangla digit and Basic character, which is a subset of Bangla character set. Different methods of sub-image or local region creation such as random creation or based on the Center of Gravity (CG) of the foreground pixels are also discussed here. Longest run features, extracted from the generated local regions, are used as local feature in the present chapter. Based on these extracted local features, together with global features, the algorithms are applied to search for the optimal set of local regions. The obtained results are higher than that results obtained without optimization on the same data set.
Chapter Preview
Top

Introduction

For recognizing different patterns identification of local regions where the pattern classes differ significantly is an inherent ability of human cognitive system. This inherent ability of human being may be imitated in any pattern recognition system by incorporating the ability of locating the regions which contain the maximum discriminating information among the pattern classes. The simplest way to do this is to divide the pattern image into a fixed number of equal sized regions. These regions may have some overlap with each other. For each such region, features (often called local features) are extracted. These local regions are then sampled randomly to produce various subsets of them. The recognition performance is evaluated with feature set formed with the local features (for some cases along with some global features)of each of those subsets. The subset, which produces best result, may be considered as an optimal set of local regions where the pattern classes differ significantly. Handwritten character recognition is a typical example of a real world pattern recognition problem which requires huge computation and modelling of perceptual power or cognitive capabilities of human beings, at least to some extent. (Cheriet, El Yacoubi, Fujisawa, Lopresti, & Lorette, 2009). To recognize the handwritten characters, both global and local features jointly or individually are used with a standard classifier. Global features are those features which are extracted from the overall character images. On the other hand, local features are extracted from the sub images of the same.

Complete Chapter List

Search this Book:
Reset