Certain improvements in classification algorithms for content based image retrieval

Abstract

Based on a given input query image Content Based Image Retrieval retrieves similar images from a large database A conventional keyword based search was inefficient in retrieving data because of large scale digitization of images diagrams and paintings A CBIR system gets inputs and responds to image queries relying on image content through use of techniques from computer vision and image processing to interpret it It uses newlinetechniques from information retrieval and databases to locate and retrieve images suiting the query CBIR is used in medicine as it increases doctor s confidence when they make informed decisions Various methods were suggested for CBIR with low level image newlinefeatures like histogram color layout texture and image analysis in the frequency domain including Fast Fourier Transform and Wavelets Similarly classification algorithms like Naïve Bayes classifier Support Vector Machine Decision Tree induction algorithms and Neural Network based classifiers were also studied extensively Future medical information systems will play a very important role in the clinical decision making process by providing similar pathological conditions in a medical image and thus help the physician view the significant images to make a better decision CBIR has been effectively used to retrieve images from databases based on the query input which can either be an anatomical region or pathological image In this work it is proposed to newlineinvestigate CBIR on medical images obtained through various techniques including Computer Tomography and Magnetic Resonance Imaging newline newline