Image Retrieval System Based On Color Features

With the continuing informatization of modern society, it becomes a bigchallenging issue that how we can establish a high efficient,accurate and stableimage retrieval system based on contents in information searching field in newperiod. On the basis of analysing and summarizing the technologies and systemswhich are available, the thesis is going to have a deep discussion on the color-basedimage retrieval area.Firstly,the thesis makes a thorough research into the characteristic of color,including the basic technology of color,color space as well as color features,whichfinally comes up with extraction of the color histogram, color moments and otherfeatures.Meanwhile,it analyses all kinds of Distance algorithm and Coefficient ofsimilarity index and Application in algorithms of similarity.With all this finished,thethesis achieves a color-based image retrieval system iCBIR by means of the mixedprogramming method consisting of MATLAB and C#.By making comparisons between color feature extracted from record ofdatabase storage and color feature of sample image,this system is able to achievecomplete function of image retrieval,using various color feature andalgotithms,thus realizing a high efficient query of which the precision and recall isperfect.