Some studies on multifeature Analysis and enhancement for Content based image retrieval

Abstract

In computer vision and image processing research the aim is to derive newlinebetter tools that give us different perspectives on the same image allowing us newlineto understand not only its content but also its meaning and significance newlineImage processing cannot compete with the human eye in terms of accuracy newlinebut it can outperform easily in observational consistency and in its ability to newlinecarry out detailed mathematical operations Also on simple or structured newlinetasks computing solutions can be reliable consistent and inexpensive Over newlinethe years image processing research has broadened from basic pixel based newlinelow level operations to high level analyses which include the use of Artificial newlineIntelligent techniques for image interpretation and understanding These new newlinetechnologies are being developed to gain a better semantic understanding of newlineimages based on the relationship between their components context and newlinehistory if they are a part of a sequence and a priori knowledge is gained from newlinea range of sources newlineIn the case of Content Based Image Retrieval CBIR it is the content newlinethat defines the object category In a majority of image analysis applications newlineobject recognition in images is a fundamental step Such research can be newlineeasily categorized as that dealing with well defined objects and that newlinedealing with natural objects Well defined objects are mostly found in newlineresearch dealing with industrial applications of machine vision and automated newlinetarget detection whereas natural objects are dealt with in remote sensing and newlineother applications newline newline