Visual information retrieval from annotated large audiovisual assets based on user profiling and collaborative recommendations

Current multimedia databases contain a wealth of information in the form of audiovisual and text data. Even though efficient search algorithms have been developed for either media, there still exists the need for abstract data presentation and summarization. Moreover, retrieval systems should be capable of providing the user with additional information related to the specific subject of the query, as well as suggest other, possibly interesting topics. In this paper, we present a number of solutions to these issues, giving an integrated architecture as an example, along with notions that can be smoothly integrated in MPEG-7 compatible multimedia database systems. Initially, video sequences are segmented into shots and they are classified in a number of predetermined categories, which are used as a basis for user profiles, enhanced by relevance feedback. Moreover, this clustering scheme assists the notion of "lateral" links that enable the user retrieve data of similar nature or content to those already returned. In addition to this, the system is able to "predict" information that is possibly relevant to specific users and present it along with the returned results.