Virage versus the BLOB: visual information retrieval technology drives a search engine. (PhotoDisk online photo archive utilizes VIR Engine image browser) (Product Information)

Visual information retrieval technology drives a search engine Quality text-searching tools are now fairly common, but it is still difficult to perform equally reliable searches of images, video and other multimedia formats. For managers of large Web sites or online archives, the new combination of multimedia content and database storage systems is a potential source of trouble, as developers and users have to work harder to keep track of their content. The database developers have made it easier to store multimedia content, but retrieval by content queries is a much harder problem. For images, content retrieval means developing visual information retrieval (vir) systems that can actually look inside an image file and search its content without relying upon keywords or other preindexed information.

One company working on a vir engine, Virage, enters the game with experience. Its founder, Dr. Ramesh Jain, helped inspire one of the first vir systems, IBM's QBIC (see The Seybold Report on Desktop Publishing, Vol. 9, No. 1). Jain founded Virage two and a half years ago to develop and market a commercial vir system. In January, it began shipping the search engine and user interface. Since then, database companies such as Informix and Object Design have licensed the Virage vir, which is operating-system independent and designed specifically for integration with oem products. For instance, Virage has developed a vir DataBlade for Illustra's object-relational database.

Paul Lego, Virage's ceo, says that the market for vir systems has developed rapidly in recent months. Cheaper disk …