Indexing and Searching Video Sequences Using Concept Lattices

The explosion of multimedia content in numerical databases have created a world need to new paradigms and techniques allowing to express how to explore, seek and summarize media collections and exploiting their content. In this paper we present a FCA-based approach to classify and search relevant video sequences from a set of video sequences. It consists in building a lattice from the binary relation between video sequences and their annotations. These annotations are extracted from a set of XML files associated to sequences. An indexation technique based on MPEG7 norm and XML language is used to annotate video sequences. Then we transform an user request into concept and merge it in the lattice. As result we obtain a set of sequences belonging to the extents of the request formel concept subsumers in the resulting lattice.