Welcome to D-MIR Lab

The mission of the Delft Multimedia Information Retrieval (D-MIR) Lab is to provide efficient, natural and intuitive access to large multimedia data collections, in a wide range of applications. The search and retrieval solutions the Lab strives for are robust, generic, computationally efficient and scalable enough to enable real-life applications and use cases. The Lab also pushes to remain at the forefront of new societal and economic developments related to digital media by making the social indexing concepts become a key ingredient of its new convergent research agenda in the MIR field.

We aim at identifying the best possibilities to combine MIR tools, such as multimedia content analysis, multimedia search reranking, query expansion and query performance prediction, with social indexing concepts, such as tagging and collaborative rating, in a user-centered fashion to optimize access to multimedia content for each individual user. We refer to this interlocking combination as the Delft Triple Synergy approach to MIR.

The work being carried out in our lab encompasses a broad palette of research directions including:

• Multimedia content analysis and search
  • Semantics extraction from multimedia data
  • Multi-modal query expansion
  • Query performance prediction
  • Multi-source search result reranking
• Multimedia information retrieval in a social network context
  • Modeling information transfer, friendship, trust and user-item-tag relations in social networks
  • Collaborative multimedia content recommendation
• User interaction for multimedia content access
  • (Affective) User profiling
  • Relevance feedback
  • Query formulation
• Content management
  • Multimedia databases and dataspaces
  • Entity retrieval

More information about the research directions and activities of the Lab can be found under research projects or on the sites of lab members.