Hybrid genetic optimized elman network based medical image classification system

Abstract

Medical devices today extensively generate large quantities of images Major challenges involved in the management of these images are indexing and classifying the images for future use Most picture archiving and retrieval systems use textual information to retrieve these images which are ineffective as most of the time the required images not being retrieved An emerging area of research is content based image retrieval where the query parameter for retrieving an image is an image Based on the query image similar images are retrieved from the database In medical imaging, automatic classification and retrieval is useful to insert the new radiographs into existing archive without interaction searching for a specific diagnoses based on an image input Image retrieval system reduces the cost in medical care significantly as the clinical decision process by a physician can be faster as anatomical features or pathologic appearance can be compared in the image database This research addresses the problem of retrieving medical images from a multi varied database The prime focus of this research is to investigate the efficacy of the various classification algorithms used in content based image retrieval and proposes a novel classification algorithm Hybrid Genetic Algorithm Optimized Elman HGAOE Neural Network to improve the classification accuracy