UB - University at Buffalo, The State University of New York

Multimedia Information Retrieval

This course will cover both traditional, text-based information retrieval (IR) techniques, as well as newer techniques for multimedia information retrieval. Although text-based IR is a well-established field, the area of multimedia IR is an emerging discipline. Techniques for automatically analyzing/extracting content from other modalities such as images, video and speech will be explored. In each case, we will discuss content-analysis techniques as well as indexing and retrieval techniques will be examined. Various case studies will be presented. The latter part of the course will focus on techniques for combining information obtained from various media sources in both the indexing and retrieval stages. This includes multimodal query formulation/ decomposition, relevance judgment/feedback, evaluation techniques, as well as convenient user interfaces for multimedia querying.

None presently available.

Basic college mathematics (calculus, linear algebra, etc.). a graduate-level course in Artificial Intelligence; preliminary image processing and/or computer vision knowledge (e.g. edge detection, pixel grouping, etc.), and NLP basics (e.g., syntax, semantics) would be helpful, but not essential.