Spatial and temporal reasoning in multimedia information retrieval and composition with XDD

This research first proposes a general framework for retrieval and composition of spatial and temporal multimedia. It is based on an equivalent transformation theory and an XDD theory. It employs the MPEG-7 and SMIL standards which provide the tools to annotate the spatial and temporal relations inside multimedia and the tools to compose the multimedia, respectively. SMIL is selected to be one of the representation languages for the multimedia documents and presentations. Yet this has no enough efficiency to handle higher-level spatial and temporal relations, whereas the MPEG- 7 provides tools to annotate those kinds of relations. These tools are used to decompose multimedia contents into segments and synchronize them together. Nevertheless, they should have more expressiveness power for retrieval and composition. As a result, the XDD theory is employed to enhance the language in terms of expressibility and computability. Moreover, by making use of XDD, constraints and axiom specifications are easily defined. This paper also defines a modeling language called SMILD that be able to represent spatial and temporal multimedia.