Skip to Main Content
The global color histogram is a well-known as a simple and often effective way to perform color-based image retrieval. However, it lacks information about the location of the image colors. While on the one hand the use of a grid of cells superimposed on the images and the use of local color histograms for each such cell improves retrieval, in the sense that some notion of color location is taken into account, on the other hand retrieval now becomes sensitive to image rotation and translation. In this short paper we present a new way to model image similarity, also using colors and a superimposing grid, via bipartite graphs. As a result the technique is able to take advantage of color location but is no longer sensitive to rotation and translation.