School of Electrical and Computer Engineering
Communications and Signal Processing Area
Video and Image Processing Laboratory (VIPER)
and
Center for Computational Image Analysis and Data Visualization
"Multiresolution-Based Image and Video Compression and Communications"
Professor Kannan Ramchandran
Beckman Institute for Advance Science and Technology
and
Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign
March 24, 1998
3:00-4:00pm
Room 117 EE Building
Internet video, video-on-demand, and wireless video are but a few examples of
the recently exploding demand for efficient and reliable image and video
communication systems. In this talk, we summarize our recent work in this
area, emphasizing the role of multiresolution representations for both source
and channel coding. We begin by highlighting recent advances in wavelet-based
image coding, and describe our contribution to the current state-of-the-art
involving the use of a variety of radically differing data-structures
-- based on zerotrees, morphological region-growing, and statistical wavelet
models -- that all share the attribute of efficiently capturing the
space-frequency characterization of the wavelet decomposition. The objective
performance of our algorithms built around these data-structures is among the
very best in the coding literature. We then highlight our work on reliable
image transmission over noisy channels based on multiresolution-based joint
source-channel coding, emphasizing the goal of maximizing the end-to-end
system performance that can lead to significant gains in delivered image
quality. Time-permitting, we will discuss the role of one or more novel
source-channel coding paradigms based on embedded modulation, arithmetic
coding based continuous error detection, hybrid compressed/uncompressed
systems, and multiple description source representations.
Professor Ramchandran is currently an Assistant Professor of Electrical and Computer Engineering and a member of the Beckman Institute for Advance Science and Technology at the University of Illinois at Urbana-Champaign. He received his Ph.D. in 1993 from Columbia University.
Professor Ramchandran is being hosted by the Video and Image Processing Laboratory (VIPER) of the School of Electrical and Computer Engineering. Please contact Prof. Edward Delp (ace@ecn.purdue.edu), 494-1740 for more information.