Courses in summer term 2009 / Seminar ML for Media Information Retrieval
Lecture | |
Time: | We 14-16 |
Location: | B26 |
Begin: | 15/04/2009 |
Assignment: | Informatic/ Ki+Ml(BSc) |
Delivery | |
Time: | |
Location: | |
Vorbesprechung: |
Mo., 2. Februar, 16 ct in B26 |
Multimedia Information Retrieval (MIR) is the problem of searching
for digital multimedia, like images, video, audio (music) in large
databases. Searching is "content-based", meaning that the search will
analyze the actual contents of the multimedia type (e.g., in
image/video it can be colors, shapes, textures). There is an enormous
growing interest in MIR, because of the large number of possible
applications. The latest trend, both commercially and in research, is
the use of social information in the form of textual description
(e.g., social tags) about images, video, or audio. Notable examples
are Flickr, YouTube, and Last.fm. Scalability to very large databases
is also an issue that has attracted significant attention, as this
requirement is important for commercial application. Such issues will
be examined in this seminar, placing emphasis on the role of machine
learning/data mining technologies in order to facilitate MIR.
Instructor: Alexandros Nanopoulos