A perspective view on visual information retrieval systems

Visual information retrieval is a new subject of research in information technology. Its purpose is to retrieve, from a database, images or image sequences that are relevant to a query. It is an extension of traditional information retrieval so to include visual media. Interactivity with visual content is essential to visual information retrieval. Visual elements such as color, texture and object shape, spatial relationship-directly related to perceptual aspects of image content, together with high-level concepts-the meaning of objects and scenes in the images, are used as clues for retrieving images with similar content from a database. The need for visual information retrieval has become apparent recently, as more and more visual information has become available in digital archives. The emergence of multimedia as the "new" technology and the possibility of sharing and distributing image/video data through large-bandwidth computer networks have made evident the importance and need of tools for retrieving visual information. Most visual information retrieval systems and experiences are prototypes that have been developed in university laboratories or research institutes. Very few commercial products are available, offering limited capabilities. Effectiveness of most solutions has still to be verified in real application contexts. Despite the fact that an increasing number of researchers are attracted by this research subject, many unexplored areas still exist and research is in its very early infancy. The paper discusses research directions in visual information retrieval.