Abstract

When we search for images in multimedia documents, we often have in mind specific image types that we are interested in; examples are photographs, graphics, maps, cartoons, portraits of people, and so on. This paper describes an automated system that classifies Web images as photographs or graphics, based on their content. The system first submits the images into some tests, which look at the image content, and then feeds the results of those tests into a classifier. The classifier is built using learning techniques, which take advantage of the vast amount of training data that is available on the Web. Text associated with an image can be used to further improve the accuracy of the classification. The system is used as a part of WebSeer, an image search engine for the Web. 1 Introduction Collections of multimedia documents can contain a vast amount of textual and visual information. However, the bigger the size of such collections grows, the harder it gets to locate specific informat...

Citations