An Improved Classification Algorithm for Content Based Image Retrieval using Fuzzy and Genetic Optimization

Abstract

Medical information systems are going to play an important role in clinically related decisions in future by rendering similar disease or pathological effects in a medical image and thereby helping physicians view noteworthy images to form improved decisions. For retrieval of images CBIR is used successfully from the databases with respect to an input query, which could be in an compartmentalize the human body image or a behavior of the pathological image. Methods suggested for CBIR include features of description of images with respect to low intensity level like histogram, texture, color, shape, and frequency domain analysis. Classifiers include algorithms such as Support Vector Machine, Neural Network, Naïve Bayes classifier and Decision Tree algorithm. In this work, it is proposed to image features can be extracted by using DCT, to extract similar features using the characteristics. For classification algorithm, Neural Network with feed forward algorithm is proposed. The optimization of learning rate by using Genetic Optimization. The proposed method of classifier is simple and it can be implemented easily in discrete frequency domain. This proposed classifier is best suited for extracted features and the computation process is for CBIR systems.