Applying experiences of providing effective information identification and retrieval for predominantly text based distributed databases to multimedia databases

This paper outlines the information management framework which has been used to design and implement a number of large scale distributed databases which contain multimedia information but where the items to be retrieved are predominantly textual and searching is associated with textual metadata. The reasons for using metadata for information management are explained as are the various means by which metadata have been obtained and augmented. Finally, the applicability of these experiences to predominantly multimedia databases is discussed.