Time: Tuesday, November 20, 10:00
Title and abstract for the talk:
Efficient adaptive retrieval and mining in large multimedia databases
There is tremendous growth in multimedia data in application domains such as medicine, engineering, biology and numerous others. Content-based similarity search and knowledge extraction are required to access and explore this data in a meaningful way. Similarity models should reflect application needs to ensure effectiveness of retrieval and mining. Moreover, as multimedia databases are typically large and high-dimensional, efficiency is a crucial aspect. We sketch some of the major challenges in retrieval and mining for different types of data. Novel approaches for efficient and effective database access are presented. Experiments demonstrate quality improvements as well as superior runtime performance compared to existing approaches.
Ira Assent is currently a Ph.D. student and research assistant at RWTH Aachen University, Germany. Her research interests are in efficient similarity search and subspace clustering in large multimedia databases. In 2003 she received her Diplom (equivalent to M.Sc.) in computer science from RWTH Aachen University.