| The World Wide Web (WWW) contains a large quantity of images and other visual information such as videos, movies, and comic strips. These images either belong to structured collections (e.g., museum collections) or are independent (e.g., images found in Web pages in the form of individual photographs, logos, and so on). The great paradox of the Web is that the difficulty in locating accurate and relevant information about a subject increases with the amount of available information. To help mitigate this problem, many information retrieval engines (or systems) have been proposed and implemented. The need for such tools is clearly shown by the success they have achieved. However, despite the visual nature of the Web, only a few engines have focused on retrieving visual information such as images and videos. Indeed, while there has been tremendous success in developing search engines for text, search engines for other media on the Web (e.g., images, audio, and video) are still rare and not as powerful. In this dissertation, we are primarily concerned with the retrieval of visual information from the Web. Visual information is published both embedded in Web documents and as stand-alone objects. It exists in the form of images, graphics, bitmaps, animations, and videos. One of the expressive and powerful languages for WWW images is XML-based Scalable Vector Graphics (SVG). SVG markup language is useful in many applications including law enforcement, image copyright protection, filtering of inappropriately mature content, criminal tracking, home entertainment, education, and training.; SVG is becoming increasingly important for displaying and exchanging multimedia data over the Internet. Although SVG is powerful in the exchange, transformation, and rendering of image data, it still has the weakness of content-based component-sensitive image search mainly because it does not yet provide the descriptors of component image objects and their spatial relationships. Additionally, although similarity matches have been studied for a long time, the similarity measures of images on the component images and their relationships are not satisfactory. |
