Performance improvement of fuzzy-based algorithms for medical image retrieval

Abstract

In this paper we present an algorithm for medical image retrieval system examines in particular image processing algorithms used for image retrieval based on transformation domain, segmentation and fuzzy logic. This is to implement image retrieval algorithm to achieve better classification accuracy for retrieving medical images. It is proposed to implement image segmentation using fuzzy logic with certain improvements for feature extraction with global optimisation using fast Hartley transform (FHT) to reduce the dimensionality of feature space. A new image segmentation algorithm, fuzzy edge detection and segmentation (FEDS) and a bell fuzzy multilayer perceptron (BF-MLP) for classifying images are used. Then BF-MLP is optimised using genetic algorithm (GA) for improved performance.