Friday, October 25, 2002
4:30-5:30 pm
1003 EECS
Abstract -
Image compression has reached the peak of efficiency in terms of
treating images as traditional "signals," employing efficient
transformations, correlation-based models, and context-based coding.
Human visual system characteristics have been successfully applied
to high-rate signal-based compression, where stimuli such as
compression-induced distortions are below the visibility threshold;
these distortions and this compression regime are called
subthreshold. Operation of such signal-based compression algorithms
in the suprathreshold regime (i.e., low-rate compression), in which
compression-induced distortions are clearly visible, has to date
operated based on visual system rules-of-thumb and has produced
moderate success. However, a unifying approach to produce
higher-quality low-rate images is lacking. In this talk I will
describe our work toward a robust, unified approach to
psychovisually motivated low-rate image compression. Experimentation
and results have produced a robust distortion allocation strategy
based on RMS contrast which is applicable to wavelet-based
compression and can be used successfully over the entire range of
compression ratios.
Biosketch -
Sheila S. Hemami received the B.S.E.E. from the University of
Michigan in 1990, the M.S.E.E. from Stanford University in 1992, and
the Ph.D. from Stanford University in 1994. She was with
Hewlett-Packard Laboratories in Palo Alto, California in 1994. In
1995 she joined the School of Electrical Engineering at Cornell
University in Ithaca, New York, where she is currently an Associate
Professor and directs the Visual Communications Lab. Dr. Hemami
received a National Science Foundation Early Career Development
Award in 1997 and has received numerous teaching awards. She held
the Kodak Term Professorship of Electrical Engineering at Cornell
University from 1996-1999, and she was a Fulbright Distinguished
Lecturer in 2001.