SEMINAR

New Directions in Image Compression

Department of Mathematics
University of Southern Mississippi

ABSTRACT

We examine an alternative approach to the current technologies being pursued in image compression used in conjunction with computer visualization and enhancement. The approach to image compression is based on functional approximation utilizing properties of probabilistic integral kernel approximations (e.g., the Bernstein functions) combined with frequency shift algorithms. These approximations can be viewed in the context of mollifyers, and subsumed within the theory of convolution operators; and, thus are related to other recent approaches to image compression using convolution. These algorithms have several desirable characteristics which naturally enhance their usability in a modern computing environments: They are adaptive and highly parallelizable.

WHERE: TEC 205

WHEN(day): Friday, Setpember 29th, 2000

WHEN(time): 2:00pm

EVERYBODY IS INVITED