Content-based Image Retrieval System by Using Color Homogram Filter and Spatial Filter

Abstract

Content-based image retrieval (CBIR) system retrieves relevant images based on their contents such as color, texture, shape, volume, spatial constraints and etc. In this paper, a CBIR system is implemented by using color homogram filter and spatial filter to improve the image retrieval system according to the image visual content by computer vision. The color feature and spatial feature information are retrieved in offline feature extraction. The color feature of an image is retrieved by using the color homogram of a CIE Lab color space and the spatial feature of an image is retrieved by the proposed spatial algorithm. A CBIR system is created by using color homogram filter and spatial filter sequentially in online query processing. The color-spatial query results from the final query ranking are ranked according to the total normalized distance in color and spatial features as the query presentation. This system is implemented by using MATLAB programming.