IEEE Circuits and Systems Society Joint Chapter of the Vancouver/Victoria Sections

Networked Multiple Description Estimation and

Compression with Resource Scalability

Speaker: Dr. Xiaolin Wu

Department of Electrical & Computer Engineering

McMaster University

Dates and Locations

Friday, February 1, 2008, 3:00 pm to 4:00 pm

Room: ASB 9896 (SFU Map)

Simon Fraser University, Burnaby, BC

Refreshment will be served after the talk.

Abstract

We present a joint source-channel multiple description (JSC-MD) framework for resource-constrained network communications (e.g., sensor networks), in which one or many deprived encoders communicate a Markov source against bit errors and erasure errors to many heterogeneous decoders, some powerful and some deprived.  To keep the encoder complexity at minimum, the source is coded into $K$ descriptions by a simple multiple description quantizer (MDQ) with neither entropy nor channel coding.  The code diversity of MDQ and the path diversity of the network are exploited by decoders to correct transmission errors and improve coding efficiency.  A key design objective is resource scalability: powerful nodes in the network can perform JSC-MD distributed estimation/decoding under the criteria of maximum a posteriori probability (MAP) or minimum mean-square error (MMSE), while primitive nodes resort to simpler MD decoding, all working with the same MDQ code.  The application of JSC-MD to distributed estimation of hidden Markov models in a sensor network is demonstrated.

The proposed JSC-MD MAP estimator is an algorithm of the longest path in a weighted directed acyclic graph, while the JSC-MD MMSE decoder is an extension of the well-known forward-backward algorithm to multiple descriptions. Both algorithms simultaneously exploit the source memory, the redundancy of the fixed-rate MDQ, and the inter-description correlations.  They outperform the existing hard-decision MDQ decoders by large margins (up to 8dB).  For Gaussian Markov sources, the complexity of JSC-MD distributed MAP sequence estimation can be made as low as that of typical single description Viterbi-type algorithms.

Biography of Dr. Xiaolin Wu

Xiaolin Wu received his B.Sc. from Wuhan University, China in 1982, and Ph.D. from University of Calgary, Canada in 1988.  He is currently a professor at the Department of Electrical & Computer Engineering, McMaster University, Canada, and holds the NSERC-DALSA research chair in Digital Cinema.  His research interests include multimedia coding and communications, image processing, signal quantization and compression, and joint source-channel coding.  He has published over one hundred sixty research papers and holds two patents in these fields.  He is the principal inventor of CALIC, the benchmark algorithm for lossless image coding.  His awards include 2003 Nokia Research Fellowship, 2000 Monsteds Fellowship, and 1998 UWO Distinguished Research Professorship.  He is an associated editor of IEEE Transactions on Multimedia, and served on program committees of numerous IEEE conferences/workshops on multimedia, data compression, and information theory.

Webcast for the talk

Please click here after 2:45pm on February 1, 2008 to join the WebEx-based webcast for the lecture. The password is 08caswu.

Please click here for a quick guide for WebEx (written by Prof. Jim Cavers).

Please note that due to technical difficulties, the webcast will have to use the SFU wireless network instead of wired one. If we could not establish the webcast because of weak wireless signal in the room, we will record the lecture using WebEx software and post it on this website later.