Signal Compression and Classification Group

Image Systems Engineering Program
The Signal Compression and Classification Group of the Information Systems Laboratory in the Department of Electrical Engineering of the School of Engineering of Stanford University does research on the theory, algorithm design, simulation, and quality evaluation of systems for signal compression, quantization, source coding with a fidelity criterion, and classification. Related topics in signal processing are also considered, including enhancement, segmentation, regression, detection, estimation, prediction, video coding, and motion estimation. A common theme of the work is the interplay of compression and classification, especially of tree-structured algorithms and clustering for vector quantization for simultaneous compression and classification.

Compression and Signal Processing Links

Location

The group resides in the Information Systems Laboratory in the Packard Electrical Engineering Building

Faculty

Robert M. Gray, PhD
Professor and Vice Chair of Electrical Engineering
rmgray@stanford.edu
Richard A. Olshen, PhD
Professor of Biostatistics and, by courtesy, Statistics and Electrical Engineering
olshen@stat.stanford.edu

Administrative Assistant

Charlotte Coe (char@isl)
Administrative support and Coordinator of the ISL and Solid-State Industrial Affiliates Programs

Students

Bradley James Betts, Ph.D. student (betts@isl): Image classification, image compression
Jia Li, Ph.D. student (jiali@isl): image processing, mainly compression
Sanjeev Mehrotra, Ph.D. student (mehrotra@voyager0): Image compression, hierarchical and finite-state codes.
Anuradha Aiyer , Ph.D. student (anu@isl.stanford.edu): Image compression

Alumni/Alumnae Association

Barry D. Andrews, 8x8 Inc. (andrews@isl): Video compression
Ender Ayanoglu (ender@research.att.com): Member of Technical Staff, AT&T Bell Laboratories.
Source and channel coding applications, communication networks.
Richard Baker (bake@pictel.com): Formerly Chief Scientist, PictureTel
Phil Chou (pachou@microsoft.com): Video compression
Pamela C. Cosman (pcosman@ece.ucsd.edu): Asst. Prof., UCSD, Data compression and image processing
Pao-Chi Chang pcchang@roger.ee.ncu.edu.tw: video/audio compression over high speed networks and wireless communications
Michelle Effros (effros@systems.caltech.edu): Asst. Prof., CalTech, Universal and adaptive data compression algorithms and theory.
Robert J. Fontana (mssi@his.com): Extremely wide bandwidth communications systems using ultra wideband, or impulse, technologies.
Sheila S. Hemami, (hemami@isl): Asst. Prof., Cornell University. Visual communications. (Very nifty site!!)
Earl Levine , (earl@isl.stanford.edu): Audio compression. Liquid Audio
Jim Kramer (kramer@virtex.com),: Virtual Technologies, Inc.
Tom Lookabaugh, (tdl@dvi.com),: Vice President Research and Business Development, Divicom
Mari Ostendorf (mo@raven.bu.edu): Assoc. Prof., Electrical, Computer, and Systems Engineering, Boston University. Data compression and statistical pattern recognition, especially speech recognition.
Keren O. Perlmutter, (keren@isl),: Senior Research Engineer, AOL, Inc; Video compression.
Sharon M. Perlmutter, (sharonp@isl),: Senior Research Engineer, AOL, Inc; Video compression.
Eve A. Riskin (riskin@rcs.ee.washington.edu): Assoc. Prof., University of Washington, Data compression and image processing. The Data Compression Group site has C-code available for downloading for full search and tree structured vector quantization.
Chieh Tsao: Formerly with the Defence Intelligence Department of the Government of Singapore. Signal processing and an award winning composer. Died of liver cancer, 1996.
Ping Wah Wong (pwong@hplpww.hpl.hp.com),: Hewlett-Packard Laboratories, image and signal processing, signal compression, digital halftoning.

Digital Mammography Image Quality Project

Former focus of the group was a project for measuring quality and utility of digital mammograms and lossy compressed digital mammograms. This project was completed in summer 1998 with the graduation of Bradley Betts, who is now with the Stanford Medical School.

Selected Group Publications

Some are postscript files, but more recent files are Adobe portable document format (pdf) 3.0.

Quantization, by R.M. Gray and D.L. Neuhoff, invited paper for the IEEE Transactions on Information Theory 50th anniversary issue (October 1998).
Introduction to Statistical Signal Processing, by R.M. Gray and L.D. Davisson.
Text and picture segmentation by the distribution analysis of wavelet coefficients, by Jia Li and Robert M. Gray, 1998 ICIP.
Context based multiscale classification of images, by Jia Li and Robert M. Gray, 1998 ICIP.
Management and legion detection effects of lossy image compression on digitized mammography", B.J. Betts, A. Aiyer, J. Li, D. Ikeda, R. Birdwell, R.M. Gray, and R.A. Olshen, Proceedings 1998 International Workshop on Digital Mammography.

Vector Quantization and Density Estimation, by R.M. Gray and R. A. Olshen, SEQUENCES97. The transparencies are also available.
Asymptotic performance of vector quantizers with a perceptual distortion measure, Jia Li, Navin Chaddha, and R.M. Gray, to appear, IEEE Trans IT.
Fundamentals of Vector Quantization, transparencies for a half day tutorial at the University of Erlangen-Nuremburg, April 1997.
Image Compression, Classification, and Universal Coding , transparencies for an invited talk at the Erlangen Workshop on Advances in Digital Image Communication, University of Erlangen/Nuremburg, 25 April 1997.
Image Quality in Lossy Compressed Digital Mammography , transparencies for an invited talk at the International Workshop on Image and Video Coding, University of Freiburg, 28 April 1997.
Lossy Compressed Medical Images and Measurement Accuracy, Sharon M. Perlmutter, Pamela C. Cosman, Chien-Wen Tseng, Richard A. Olshen, and Robert M. Gray, King C. P. Li, and Colleen J. Bergin. To appear, Statistical Science, revised 4/10/97. Also available in postscript format.
"Image Quality in Lossy Compressed Digital Mammograms," S.M. Perlmutter, P.C. Cosman, R.M. Gray, R.A. Olshen, D. Ikeda, C.N. Adams, B.J. Betts, M. Williams, K.O. Perlmutter, J. Li, A. Aiyer, L. Fajardo, R. Birdwell, B.L. Daniel, to appear, Signal Processing, Special Section on Medical Image Compression, in portable document format (pdf) format. (Adobe Acrobat Reader 3.0). 5/97.

"Evaluating quality and utility of digital mammograms and lossy compressed digital mammograms," Proceedings of the Third International Workshop on Digital Mammography, Elsevier, Amsterdam, pages 169-176. Preprints available in either postscript or portable document format (pdf).
Toeplitz and Circulant Matices: A Review, by R. M. Gray. A very old (1971, revised 1977, 1993, 1997, 1998) but still occasionally useful tutorial on Toeplitz and circulant matrices.
Signal Processing and the International Information Infrastructure, R.M. Gray, Proceedings of the International Conference on Telecommunications, Istanbul, Turkey, April 1996. This page includes a link to a revised and updated version of the report Signal Processing and the National Information Infrastructure cited below. It also contains a collection of nifty URLs on signal processing for the internet and vice versa.
Combining signal compression with classification, regression, and segmentation Slides for a presentation at the Center for Applied Signal Processing (CRASP) in Ft. Meade, Md.
Constrained and recursive hierarchical table-lookup vector quantization , Navin Chaddha, P.A. Chou, and R.M. Gray, Proceedings 1996 DCC.
A vector quantization approach to universal noiseless coding and quantization, by P.A. Chou, M. Effros, and R.M. Gray, IEEE Transactions on Information Theory.
Text segmentation in mixed-mode images using classification trees and transform tree-structured vector quantization , K.O. Perlmutter, N. Chaddha, J.B. Buckheit, R.M. Gray, and R.A. Olshen, ICASSP 1996.
Vector quantization of image subbands: a survey, P.C. Cosman, R.M. Gray, and M. Vetterli, IEEE Transactions on Image Processing, February 1996.
Bayes risk weighted vector quantization with posterior estimation for image compression and classification , Keren O. Perlmutter, Sharon M. Perlmutter, Robert M. Gray, Richard A. Olshen, and Karen L. Oehler IEEE Transactions on Image Processing, February 1996).
Measuring Quality in Computer Processed Radiological Images, R. M. Gray, R. A. Olshen, D. Ikeda P.C. Cosman, S.M. Perlmutter, C. Nash, K.O. Perlmutter, presented at the 1995 Asilomar Conference on Circuits, Systems, and Computers. The transparencies are also available.
Quantization Noise in Delta Sigma A/D Converters, R.M. Gray. (Contributed chapter in IEEE book on Delta Sigma converters. (copyrighted by IEEE.))
Predicting high risk cholesterol levels Garber, A.M., Olshen, R.A., Zhang, H., and Venkatraman, E.S., International Statistical Review,62, 2, (1994), 203-228. This paper summarizes a variety of longitudinal models for predicting changes in cholesterol over several years so that reasonable policies for screening can be formulated. It includes bootstrap-based tests of whether measured cholesterol is Markovian (it is not) or a low order ARMA process (no again). There are several approaches to ROC curves for schemes whereby the possibility of a high risk value is predicted from past values; these include new techniques for computing bootstrap-based confidence regions for points on the curves.
Signal Processing in the National Information Infrastructure , Edited by R.M. Gray. Report of a Workshop/Panel held at the National Science Foundation, Circuits and Signal Processing Program, Microelectronics Information Processing Systems Division, Computer and Information Science and Engineering Directorate, August 1994. There is also an html version. See also the summary transparencies used in the NSF presentation.
Introduction to Statistical Signal Processing, by R.M. Gray and L.D. Davisson. This is a book in Adobe portable document format (or postscript).
Non-US Data Compression and Coding Research, R.M. Gray (Ed.), M. Cohn, L.W. Craver, A. Gersho, T. Lookabaugh, F. Pollara, and M. Vetterli, November 1993. A Foreign Applied Sciences Assessment Center (FASAC) report prepared for Science Applications International Corporation (SAIC) under U.S. Government sponsorship. This is the draft version submitted in LaTeX form to SAIC. The subsequent official Report was a further edited version of this report in Word.

Last modified: November 1999

Data Compression graphic courtesy Jim Storer, from the cover of his book Data Compression, Computer Science Press.