An infrastructure for mining medical multimedia data

Biomedical research processes related to disease diagnosis, prognosis and monitoring would great benefit from advanced tools able not exclusively to store and manage multimodal data but also to process and extract significant relations and then novel knowledge from them. Indeed, making a prediction on a disease outcome usually requires considering heterogeneous pieces of information obtained from several sources which should be compared and related. Mining medical multimedia objects is aimed at discovering and making available the hidden useful knowledge embedded in collections of data and is, then, of key importance for supporting clinical decision-making. In this paper, we report current results of a medical warehouse we are developing in an integrated environment for mining clinical data acquired by different media. In particular, focus is herein given to the infrastructure of the warehouse and its current functionalities not limited to storage and management but including intelligent representation and annotation of multimedia objects.