Fall 1995
sclaroff@cs
Office Hours: Thursday 2:00-5:00 or by appointment
This seminar will cover recent advances in organizing and searching multimedia databases based on their content, with an emphasis on representational methods for image and video databases. Specifically covered will be strategies for automated database stratification and indexing based on color, texture, shape, grayscale appearance, motion, and combinations of these measures. Other topics include: face recognition and detection, foreground/background segmentation, camera motion and scene cut detection, and video keyframe extraction. For a preliminary list of class topics, see the syllabus.
Permission of the instructor
Previous exposure to signal processing and mathematics through calculus of two variables, including matrices and linear spaces, and a familiarity with basic probability.
Seminar participants will take turns in leading class discussions about assigned readings. In addition, everyone in the seminar (auditors included) will prepare a brief weekly write-up that summarizes and critiques the week's readings.
By mid-term, each member of the seminar is expected to write a one page description of their proposed final programming project. Ideally, this project will implement and analyze a research paper, or it can utilize image/video database techniques to solve a specific problem relevant to the graduate student's own research. Completed projects will be due at the last seminar meeting, and will be evaluated based upon both an in-class presentation (or demo) and a written report.
Stan Sclaroff
Created: May 17, 1995