[Home]
[Back Issues]
[Search]
[Vol 1]
[Vol 2]
[Vol 3]
[Vol 4]
[Vol 5]
The Wavelet Digest - Volume 4, Issue 11 (November 4, 1995)
Today's Editor: Wim Sweldens
AT&T Bell Laboratories, Murray Hill
wim@research.att.com
Today's Topics:
1. Book: Signal processing with fractals: A wavelet-based approach
2. Preprint: Biorthogonal Multiwavelets
3. Preprint: Two preprints from Carl Taswell
4. Preprint: Vannucci-Vidakovic: Wavelet density estimation (4)
5. Thesis: Multiplexed Communication Via Wavelet Packets
6. Meeting: Recent Advances in PDE's and applications
7. Meeting: Winter School on Iterative Methods
8. Meeting: AMS Annual meeting, Florida
9. Meeting: Call for papers, SPIE96 Denver
10. Meeting: Course on wavelets and filter banks
11. Meeting: American Geophysical Union
12. Meeting: Wavelets Workshop on Advanced Data Processing
13. Course: Wavelet course at Umass Lowell
14. Answer: Use of Numerical Recipes software (WD 4.9 #10)
15. Answer: Parametrization of Biorthogonal Wavelets (WD 4.10 #24)
16. Question: Looking for lapped filter banks
17. Question: Fractional Brownian Noise
18. Announce: Filter banks with more than 50 significant digits
Submissions:
E-mail to wavelet@math.sc.edu with "submit" as subject.
Subscriptions:
E-mail to wavelet@math.sc.edu with "subscribe" as subject.
To unsubscribe, e-mail with "unsubscribe" followed by your e-mail
address as subject. To change address, unsubscribe and resubscribe.
Back issues and keyword :
FTP: ftp.math.sc.edu (/pub/wavelet)
WWW: http://www.math.sc.edu/~wavelet/
Current number of subscribers: 5626
Calendar of events:
Nov 9-10: Wavelets Workshop on Advanced Data Processing, Orlando
Nov 14: Tutorial on Wavelets, Tokyo
Nov 27-Dec 1: Acoustical Society of America, St. Louis
Nov 29-Dec 1: UCLA Multirate Digital Filters, LA
Dec 10-13: Neural Networks and Signal Processing, Nanjing, China
Dec 11-15: American Geophysical Union, San Francisco, CA
Dec 14-20: Winter School on Iterative Methods, Hong Kong
-- 1996 --
Jan 10-13: AMS Annual meeting, Florida
Jan 13-15: Course on wavelets and filter banks, USF, Tampa
Jan 27-Feb 2: Still-Image Compression, Photonics West, San Jose
Jan 28-Feb 2: Digital Video Compression, San Jose, CA
Apr 29-May 3: 1/f Long Range Dependent Processes, Les Houces, Fr.
Jun 10-14: Recent Advances in PDE's and applications, Venice, It
Jun 18-21: IEEE Symp. Time-Frequency/Time-Scale Analysis, Paris
Aug 4-9: SPIE, Wavelet applications, Denver
From: gww@allegro.mit.edu (Gregory Wornell)
Subject: Book: Signal processing with fractals: A wavelet-based approach
NEW BOOK
SIGNAL PROCESSING WITH FRACTALS: A WAVELET-BASED APPROACH
by Gregory W. Wornell, Massachusetts Institute of Technology
Prentice-Hall Signal Processing Series,
October 1995, 177 pp., ISBN 0-13-120999-X
Fractal geometry and recent developments in wavelet theory are having
an important impact on the field of signal processing. Efficient
representations for fractal signals based on wavelets are opening up
new applications for signal processing, and providing better solutions
to problems in existing applications. SIGNAL PROCESSING WITH FRACTALS
provides an valuable introduction to this new and exciting area, and
develops rich perspectives on the topic. Powerful techniques for
synthesizing, analyzing, and processing fractal signal for a wide
range of applications are developed in detail, and novel applications
in communications are explored. The book is self-contained, and in
particular it contains a highly accessible concept-oriented primer of
the relevant wavelet theory.
From: xu@hammilton.math.ndsu.NoDak.edu
Subject: Preprint: Biorthogonal Multiwavelets
Title: Reconstruction and Decomposition Algorithms for Biorthogonal
Multiwavelets
Authors: Charles A. Micchelli and Yuesheng Xu
Abstract: We construct biorthogonal multiwavelets in a weighted Hilbert space
$L^2(E,\rho)$, where $E$ is a compact subset in $R^d$.
To get a copy of the preprint, send email message to xu@plains.nodak.edu
From: "Carl Taswell" <taswell@wavbox.com>
Subject: Preprint: Two preprints from Carl Taswell
The following preprints are available for downloading from the web at
http://www.wavbox.com/wbpapers.html; please send any comments to the
author; helpful criticisms and suggestions appreciated.
Taswell, C. "Speech Compression with Cosine and Wavelet Packet
Near-Best Bases", Scientific Computing and Computational Mathematics,
Stanford University, August 1995.
Compression of speech from the TIMIT corpus was investigated for
several transform domain methods coding near-best and best bases from
cosine and wavelet packet transforms. Satisficing (suboptimizing)
search algorithms for selecting near-best bases were compared with
optimizing algorithms for best bases in these adaptive tree-structured
transforms. Experiments were performed on several hundred seconds of
speech spoken by both male and female speakers from all dialect
regions of the TIMIT corpus. Near-best bases provided rate- distortion
performance effectively as good as that of best bases but without the
additional computational penalty. Cosine packet bases outperformed
wavelet packet bases.
Taswell, C. "Algorithms for the Generation of Daubechies Orthogonal
Least Asymmetric Wavelets and the Computation of Their Holder
Regularity", Scientific Computing and Computational Mathematics,
Stanford University, August 1995.
Explicit algorithms are presented for the generation of Daubechies
compact orthogonal least asymmetric wavelet filter coefficients and
the computation of their Holder regularity. The algorithms yield
results for any number $N$ of vanishing moments for the
wavelets. These results extend beyond order $N=10$ those produced by
Daubechies for the values of the filter coefficients and those
produced by Rioul for the values of their Holder regularity. Moreover,
they reveal that the choice of phase for the filters published by
Daubechies for orders $N=4$ to $N=10$ was not made consistently. In
particular, her filter coefficients for orders $N=7$ to $N=9$ should
be reflected to their mirror image sequence.
Carl Taswell, MD, PhD Tel: 415-323-4336
WavBox ToolSmiths Fax: 415-323-5779
Post Office Box 9925 Email: taswell@wavbox.com
Stanford, CA 94309-9925 WWW URL: http://www.wavbox.com/
WavBox Software; ToolSmiths Software and Consulting;
WavBox and ToolSmiths are trademarks of WavBox ToolSmiths.
From: Brani Vidakovic <brani@DAUB4.isds.duke.edu>
Marina Vannucci <vannucci@isds.duke.edu>
Subject: Preprint: Vannucci-Vidakovic: Wavelet density estimation (4)
Four Manuscripts are available.
http://www.isds.duke.edu/ or ftp.isds.du
Any comment is welcome.
1. "Nonparametric Density Estimation using Wavelets "
Marina Vannucci University of Florence Italy
This paper addresses the problem of density estimation using the
Daubechies wavelets. Wavelet density estimators have been proposed by
several authors. It has been shown that the local nature of wavelet
functions makes these estimators competitive with classical projection
estimators. Further, the problem of negative estimates can be
rectified by estimating the square root of the density. This paper
starts with a review of classical and wavelet density estimators. Then
an extension of the univariate to the multivariate case is
presented. Finally, an application to a bivariate dataset is
implemented.
File: ftp.isds.du/pub/WorkingPapers/95-26.ps
--
2. "PREVENTING THE DIRAC DISASTER:Wavelet Based Density Estimation"
Marina Vannucci University of Florence Italy
Brani Vidakovic Duke University USA
This paper addresses the problem of optimal number of basis functions
in the wavelet series estimator of density. It is a well known that
projection estimators tend to overfit the density if the number of
basis functions in the orthogonal expansion is too large. In extreme
cases the estimator is close to the Dirac function concentrated at the
observations. By exploiting the idea of Good and Gaskins (1971) an
objective way of determining the number of levels in the wavelet
density estimate using the Fisher information functional as a
roughness penalty measure is proposed. The method is exemplified on
the galaxy and old faithful geyser data sets as well as on some
simulated data.
File: ftp.isds.du/pub/Users/brani/papers/WavDensFisher.ps
or ftp.isds.du/pub/WorkingPapers/95-27.ps (264Kb)
--
3. On wavelet scalograms and their applications in economic time series
By Miguel A. Ari\~{n}o} Universidad de Navarra
and Brani Vidakovic
Duke University
{\bf Abstract.} The scalogram is the discrete wavelet transformation (DWT) counterpart
to the well-known notion of periodogram in the spectral
analysis of time series. In the same manner as the periodogram
produces an ANOVA decomposition of energy of a signal to different
Fourier frequencies, the scalogram decomposes the energy to
``level components.''
In this paper we show how DWT and scalograms can be used in
detecting and separating periodic components in time series.
The proposed method is exemplified with a Spanish cement
production data set.
\noindent{\bf Key words and phrases:} Wavelets, Time series, Scalograms.
File: ftp.isds.du/pub/Users/brani/papers/WavTS.ps 353 Kb
--
4. Bayesian Inference with Wavelets: Density Estimation
{\large {\sc Peter M\"uller} and {\sc Brani Vidakovic}}
\end{center}
\vspace{.5cm}
\subsection*{Abstract}
We propose a prior probability model in the wavelet coefficient
space. The proposed model approaches wavelet coefficient thresholding
by full posterior inference in a coherent probability
model. We introduce a prior probability model with mixture
priors for the wavelet coefficients $d_{jk}$. The prior includes
a positive prior probability mass for zero coefficients. This
naturally leads to non-linear a posteriori shrinkage and thresholding
on the coefficients. We discuss an efficient posterior simulation
scheme to implement inference in the proposed model.
The discussion is focused on the density estimation problem. However,
the introduced prior probability model on the wavelet coefficient
space is general.
\vspace{.5cm}
{\it Keywords:} Wavelet decomposition; Model choice;
Density estimation; Posterior simulation.
File: ftp.isds.du/pub/Users/brani/papers/WavMCMC.ps 241 Kb
Comments welcome.
----
Note from the editor:
Check that Brani Vidakovic's machine listens to the name DAUB4 !
From: lindseya@rl.af.mil
Subject: Thesis: Multiplexed Communication Via Wavelet Packets
The following work was completed in June 1995 and is publicly
available via anonymous ftp at voyager.cns.ohiou.edu in directory
/pub/papers/lindsey/dissert.
The author can be reached at lindseya@rl.af.mil or via snail-mail at
RL/C3BB, 525 Brooks Road, Griffiss AFB, NY 13441-4505
Ph.D. DISSERTATION, June 1995:
Alan R. Lindsey
Electrical and Computer Engineering
Ohio University, Athens, Ohio
Generalized Orthogonally Multiplexed Communication Via Wavelet Packet
Bases (118pp.)
Building on recently introduced multidimensional signaling
techniques, a generalized multirate wavelet-based modulation format is
presented. Wavelet Packet Modulation employs customizable wavelet
packet basis functions from an arbitrary pruning of a dyadic filter
bank as novel pulse shapes upon which independent QAM data at lower
rates are placed; in fact, QAM is one of many special cases of this
general scheme. The advantages include dimensionality in both time and
frequency for flexible channel exploitation and an efficient and
practical transceiver implementation with all-digital filter banks.
The filter bank implementation also allows for adaptation and a method
for optimizing the supersymbol for channels corrupted by tones and
impulses is introduced. Supersymbol tuning, which isolates these
common noise sources to a fraction of the supersymbol, allows much of
the transmitted data to pass without interference. The superior
computational performance of the algorithm is illustrated in light of
the explosive combinatorics of the wavelet packet libraries, and more
complex enhancements/alternatives are given.
Subjects: Communications, Feedback Adaptive, Wavelet Transform
Implementations
Keywords: Wavelets, Communication, Time-frequency, Multi-dimensional
Signaling, Adaptation, Wavelet Packets, Orthogonally Multiplexed
Communication, Supersymbol Tuning
From: Renato Spigler <spigler%ipdudmsa.bitnet@icineca.cineca.it>
Subject: Meeting: Recent Advances in PDE's and applications
A conference entitled "Recent Advances in Partial Differential
Equations and Ap plications" will be held in Venice, Italy, between
June 10 and 14, 1996. The c onference will be dedicated to Peter
D. Lax and Louis Nirenberg on their sevent ieth Birthday. Invited
speakers will be Cercignani, Chorin, Da Prato, De Giorg i, Deift,
Gallavotti, Klainerman, Levermore, P.L.Lions, McLaughlin, Majda,
Mosco, Papanicolaou, Sanz-Serna, Tadmor, and Venakides.
For more information, send a message to
venice96 at ulam.dmsa.unipd.it
Renato Spigler
From: Raymond Chan <rchan@math.cuhk.hk>
Subject: Meeting: Winter School on Iterative Methods
Winter School on Iterative Methods in
Scientific Computing and Their Applications
Host by:
The Institute of Mathematical Sciences
The Chinese University of Hong Kong, Hong Kong
December 14--20, 1995
Main Invited Lecturers and Sample Topics:
Tony Chan (UCLA): Domain Decomposition, Image Processing.
Jack Dongarra (U. of Tennessee): Algorithms, Libraries.
Howard Elman (U. of Maryland): Stokes problems, Oseen Problems.
Gene Golub (Stanford): Error Estimation, Inner-Outer Iterations.
Franklin Luk (RPI): Signal and Image Processing.
David Silvester (UMIST): FEM, Navier-Stokes Equations.
Gilbert Strang (MIT): Wavelets, Filters.
Henk van der Vorst (U. of Utrecht): Krylov Methods, Implementation.
Andy Wathen (UK): Stokes Problems, Preconditioning.
Jinchao Xu (Penn. State): Multilevel Methods, Convergence Theory.
School Model
It will be a seven-day school with 10 main invited speakers.
Each main speaker will give three 40-min lectures.
There will also be several other lectures given by
speakers from the Asian region. There will be 5-6
lectures per day except for the fourth day which will
be reserved for excursion or social activities.
Demonstration and exercise sessions will be at the end of each day.
Lecture notes of the School will be distributed to attendees.
Call for Posters
The School will have a poster session for participants to exhibit
their research results. Authors are invited to submit their papers
(maximum four pages) on any topics related to scientific computing.
Please send the papers electronically or in hard copy form to
Dr. K.M. Yeung
Department of Mathematics
Chinese University of Hong Kong
Shatin, Hong Kong
Email: kmyeung@cuhk.hk
Deadline for submission is December 1, 1995.
Registration
Registration fee is US$100 which includes all lecture notes
to be distributed at the School. To receive the registration
form, please send request to Dr. K.M. Yeung at address above.
A latex file of the form can be obtained by anonymous ftp at
ftp://ims.cuhk.hk/conf/winter/regis.tex.
General Information
Detail information of the School can be obtained at
http://www.math.cuhk.hk/conference/dec95/info.html.
From: zalikri@duc.auburn.edu (Richard A. Zalik)
Subject: Meeting: AMS Annual meeting, Florida
Special Session on Computational Harmonic Analysis and Approximation
Theory to be held during the annual meeting of the American
Mathematical Society, Orlando, Florida, January 10 - 13, 1996.
Organizers: R. A. Zalik and N. K. Govil, Auburn University.
List of participants
A. Aldroubi (National Institutes of Health)
G. Anastassiou (Memphis)
L. Auslander (CUNY)
J. J. Benedetto (Maryland)
P. Borwein (Simon Fraser)
C. Cabrelli and U. Molter (Buenos Aires)
I. Daubechies (Princeton)
G. Davis (Dartmouth)
N. K. Govil and R. A. Zalik (Auburn)
K. Grochenig (Connecticutt)
D. Healy (Dartmouth)
M. Ismail (USF)
M. A. Kon (Boston) and L. Raphael (Howard)
D. Lubinsky (Witwatersrand)
D. Maslen (Utrecht)
M. Z. Nashed (Delaware)
R. Mohapatra (Orlando)
T. E. Olson (Dartmouth)
D. Rockmore (Dartmouth)
G. Strang (MIT)
V. Strela (MIT)
R. Tolimieri (CUNY)
M. Taylor (National Center for Atmospheric Research)
D. Walnut (George Mason)
The talks will be given on Wednesday January 10 (morning and
afternoon), and on Thursday morning.
From: unser@helix.nih.gov (Michael Unser)
Subject: Meeting: Call for papers, SPIE96 Denver
CALL FOR PAPERS (250 word summary due to SPIE by 8 January 1996)
WAVELET APPLICATIONS IN SIGNAL AND IMAGE PROCESSING IV
4-9 August 1996, Denver, Colorado, USA
Conference Chairs: Michael Unser, National Institutes of Health
Akram Aldroubi, National Institutes of Health
Andrew F. Laine, Univ. of Florida;
Program Committee:
John J. Benedetto, Univ. of Maryland/College Park;
Ingrid Daubechies, Princeton University;
Ron DeVore, Univ. of South Carolina;
Patrick Flandrin, CNRS (France);
Georgios B. Giannakis, University of Virgina, Charlottesvile;
Dennis M. Healy, Jr., Darmouth College;
Bjorn Jawerth, Univ. of South Carolina;
Arun Kumar, Southwestern Bell;
Stephane G. Mallat, New York Univ.;
Yves Meyer, University Paris-IX Dauphine;
Gilbert Strang, Massachusetts Institute of Technology;
Wim Sweldens, AT&T Bell Laboratories, Murray Hill;
Martin Vetterli, Swiss Federal Institute of Technology, Lausanne;
This conference, which is open to researchers in mathematics,
signal and image processing, computer vision, medical imaging,
and physics, shall focus on novel applications of wavelet-based
signal analysis and processing methods, refinements of existing
techniques, and new theoretical developments. One of the primary
goals will be to facilitate interdisciplinary exchange and to
generate cross-fertilization between these various fields of research.
Keynote speaker: Martin Vetterli.
Topics for the conference may include but are not limited to:
- image pyramids
- frames and overcomplete representations
- multiresolution algorithms
- wavelet-based noise reduction and restoration
- multiscale edge detection
- design of nonlinear operators
- wavelet texture analysis and segmentation
- Gabor transforms and space-frequency localization
- wavelet-based fractal analysis
- multiscale random processes
- wavelets and neural networks (pattern recognition)
- image representations from wavelet maxima or zero-crossings
- wavelet compression, coding and signal representation
- wavelet theory and multirate filterbanks
- wavelets in medical imaging.
This conference (with 80 papers in 1995) is expected to be one of the
largest of the 77 SPIE conferences to be held at the Optical Science,
Engineering, and Instrumentation symposium, Denver 4-9 August 1996.
TO OBTAIN ALL DENVER'96 CALLS FOR PAPERS ELECTRONICALLY:
- SPIE Web: http://www.spie.org/web/meetings/denver_home.html,
- by anonymous FTP: ftp://spie.org/meetings/calls/colo96_readme.txt,
- by e-mail: info-spie-request@spie.org with the following
in the message body: send [meetings.calls]colo96_readme.txt
Prospective authors are invited to submit wavelet papers in any of the
areas listed above. All submissions will be reviewed; camera ready
manuscripts will be required of all accepted applicants. These will be
published in a special volume of the Proceedings of the SPIE.
GUIDELINES FOR SUBMITTING AN ABSTRACT
Send a 250 word abstract of your paper, by 8 January 1996,
in ONE of the following ways:
- mail (please mail 4 copies) to:
DENVER96
SPIE, P.O. Box 10, Bellingham, WA 98227-0010
Shipping Address: 1000 20th Street, Bellingham, WA 98225
Telephone: 360/676-3290
- electronic mail in ASCII format to Internet abstracts@spie.org
Please submit one abstract per email message and place the
word DENVER96 on the subject line.
- fax to SPIE at 360/647-1445
Please submit one abstract per fax.
Be sure each abstract includes the following:
1. CONFERENCE CHAIR and CONFERENCE TITLE
Chairs : Unser, Aldroubi, Laine
Wavelet Applications in Signal and Image Processing IV
2. AUTHOR LISTING (List principal author first)
for each author: full name and affiliation, mailing address,
phone/fax numbers, email
3. ABSTRACT/PAPER TITLE
4. ABSTRACT TEXT: 250 words typed on white paper
5. KEYWORDS: maximum of 5 keywords
6. BRIEF BIOGRAPHY of the principal author: 50-100 words
Michael Unser
BEIP, Building 13/3N17
National Institutes of Health
Bethesda, MD 20892-5766, USA
From: Gil.Strang@comlab.oxford.ac.uk
Subject: Meeting: Course on wavelets and filter banks
FIRST ANNOUNCEMENT
WORKSHOP COURSE ON WAVELETS AND FILTER BANKS
taught by Gilbert Strang (MIT) and Truong Nguyen (Wisconsin)
Saturday-Sunday-Monday(holiday) January 13-15, 1996
University of South Florida at TAMPA
TEXT: Participants will receive the new textbook (publication date 1/1/96)
WAVELETS AND FILTER BANKS by Strang and Nguyen
Wellesley-Cambridge Press, Box 812060, Wellesley MA 02181
We will aim for the right balance in this Workshop. The text gives
an overall perspective of the field - which has grown with amazing speed
in theory and applications. The topics will include
Multirate Signal Processing: Filtering, Decimation, Polyphase
Analysis and Design of Filter Banks
Matrix Analysis: Toeplitz Matrices and MATLAB Implementation
Wavelet Transform: Pyramid and Cascade Algorithms
Daubechies Wavelets, Orthogonal and Biorthogonal Wavelets
Audio and Image Compression, Quantization Effects
Transient Detection and Radar Processing
Non-Destructive Evaluation, Digital Communication
Perfect Reconstruction: Cosine-Modulated Filter Banks
Spectral Factorization and Transmultiplexers
Time-Frequency and Time-Scale Analysis
Imaging and Scientific Visualization
The goal of the Workshop is to be as useful as possible to all
participants. Please request information by an email message
with subject Workshop to the organizer
Gilbert Strang: gs@math.mit.edu
We will reply about the program and registration and housing.
The tuition cost will be reduced by 50% for graduate students.
We are very glad to answer all questions.
From: fournier@cloudy.geology.yale.edu (Aime' Fournier)
Subject: Meeting: American Geophysical Union
American Geophysical Union 1995 Fall Meeting, December 11-15, 1995,
San Francisco, California:
Here is a chronological list of presentations that have "wavelet" in
the title; gleaned from http://www.agu.org/fm95top.html.
Tuesday:
0830h A21C-08 POSTER Wavelet Analysis of Daily Total
Ozone in Order to Identify Processes
Contributing to Short-Term Variability: *B L
Weinberg, R A Reck
0830h S21B-01 POSTER Wavelet Polarization Analysis of
Terrascope Seismic Network Data: Surface
Wave Refraction in Southern California: *L
Su, J Park
Wednesday:
1100h H31C-08 The Partitioning of Attached and Detached
Eddy Motion in the Atmospheric Surface Layer
Using Lorentz Wavelet Filtering: *G Katul, B
Vidakovic
1330h H32C-10 POSTER Application of Wavelets to Obtain
Atmospheric Boundary Layer Structural
Information From Elastic Lidar Measurements:
Y Cui, M B Parlange, A T Cahill, T A
Ortenburger, W E Eichinger
1330h O32B-24 POSTER Wavelet Spectral Analysis of the
Earth's Obliquity for Paleoclimate Cycles:
*H S Liu, B F Chao
Thursday:
0830h A41A-02 POSTER Wavelet Zonal Spectral Analysis of
Observed Geopotential and Winds:
Scale-Orthogonal Decomposition of Blocking
Patterns and Local Kinetic Energy and
Enstrophy Transfer Between Scales: *A
Fournier
1330h SM42B-31 POSTER Wavelet Analysis and Neural Network
for Plasma Wave Classification Onboard
Satellite: *T Toda, S Yagitani, I Nagano, K
Hashimoto, T Okada, H Matsumoto
Abstract received:
6196D M Gendron Development of a Seismic Event Detection and
Identification Algorithm Based on Wavelet
Transforms
Aime' Fournier
Yale University
http://pantheon.cis.yale.edu/~fournier/fournier.html
From: Charles_E_Stirman@ccmail.orl.mmc.com
Subject: Meeting: Wavelets Workshop on Advanced Data Processing
Martin Marietta Technologies, Inc. (MMTI) in conjunction with Fast
Mathematical Algorithms and Hardware (FMA&H) and ARPA are pleased to
announce its 3rd biannual Wavelets Workshop on Advanced Data
Processing for Automatic Target Recognition tasks. The workshop is
sponsored by ARPA and is to be held at the Orlando Facility on Sand
Lake Road on November 9 and 10, 1995 starting at 8:00 AM Eastern
Standard Time.
This year the featured speakers are:
Professor Ronald Coifman from Yale University;
Professor Victor Wickerhauser from Washington University in St. Louis;
Dr. Naoki Saito from Schlumberger Research;
Professor Greg Warhola from the Air Force Institute of Technology;
and Dr. Michael Tucker from Fastman Corp.
The workshop will cover the following topics:
an introduction to wavelets and their selection;
various Interactive Software Toolkits for analysis and algorithm development;
automatic feature extraction;
background and clutter removal;
modeling and denoising;
discrimination and classification;
regression or parameter estimation;
image enhancement tools;
multiscale sharpening and graphic equalization;
compression (several image and video compression schemes will be described);
estimating regularity; and
a variety Department of Defense applications.
A set of lecture notes will be provided to attendees Thursday morning.
In addition, as in previous years, the lectures will be video taped.
An optional registration fee of $250.00US payable to FMA&H will be
charged for those wishing NTSC version of the VHS video taped
lectures. Those paying the optional fee will receive the tapes at no
extra charge. For those not wishing the video tapes, no registration
fee is required. However, registration is required so we can prepare
for snacks, lunch, parking and sufficient seating (note: for those
requiring PAL format and overseas shipping, there will be an extra
charge of $150.00US for the conversion plus UPS shipping
charges). FMA&H will take American Express, Money Orders or Certified
checks made out to FMA&H.
Attendees can stay at the Embassy Suite Plaza International Hotel,
8250 Jamaican Court, Orlando, Florida. It is located 1/2 mile from the
workshop site and, based on availability, you can call them at (407)
345-8250 and book your room at the low rate of $70.00 per night.
HOWEVER, YOU MUST BOOK YOUR ROOM NO LATER THAN 31 OCTOBER TO RECEIVE
THIS RATE. After 24 October, the rate is subject to availability.
Please mention that you are attending the Wavelets Workshop at
Lockheed Martin to receive the reduced rate.
For registration please contact Ms. Sue Archer at
407-356-2583 tel
407-356-0933 fax
Susan_J_Archer@ccmail.orl.mmc.com or
Charles_E_Stirman@ccmail.orl.mmc.com
The following information is needed to arrange for entry
into the MMTI facility:
1. name
2. affiliation
3. address
4. zip or postal code
5. telephone number
6. social security number
7. date and place of birth(city, state and country)
8. fax number
9. e-mail address
10. do you wish to pay the optional registration and receive the video tapes?
yes or no. if yes, what format, NTSC or PAL
1. Hotel Accommodations: The $70.00 rate at Embassy Suites is
available until October 31. After 31 October, the rate is subject to
availability.
2. Registration: Registration is required as we are an approved DoD
facility: the deadline is November 3rd. If you wish to obtain video
copies of the workshops, there is an optional registration fee of
$250.00US payable to FMA&H for NTSC versions of VHS that will cover
mailing within the United States. (For
PAL versions of VHS there is an extra $150.00US charge. For shipment
outside the United States you need to add appropriate UPS shipping
charges. Call Fran Kerns at FMA&H (203)-248-8212 for shipping charges
outside the United States.) FMA&H will take American Express, Money
Orders or Certified Checks made out to FMA&H.
From: KAISERG@aspen.uml.edu
Subject: Course: Wavelet course at Umass Lowell
COURSE: Wavelet Analysis, 92.552
INSTRUCTOR: Dr. Gerald Kaiser, Mathematical Sciences Dept., UML
E-MAIL: kaiserg@woods.uml.edu
TEXTBOOK: G. Kaiser, A Friendly Guide to Wavelets, Birkhauser, 1994
TIME: Wednesdays 6-9PM, starting 1/17/96
PLACE: Olsen Bg., probably Rm 306, UML (North Campus)
The prerequisites are minimal: Matrix theory, Fourier series, and
Fourier integral transforms. Some experience with Fourier analysis
(as practiced, for example, by signal analysts) would help but is not
necessary. All the necessary mathematics are developed in the first
chapter of my book, which will be covered thoroughly with emphasis
on intuition rather than rigor.
The coverage will consist mainly of Part I of the book,
BASIC WAVELET ANALYSIS:
Mathematical background and notation
Continuous and discrete time-frequency analysis
Continuous and discrete time-scale (wavelet) analysis
Generalized frames: Key to analysis and synthesis
Multiresolution analysis and subband filtering
DaubechiesU wavelet bases.
If time permits and there is sufficient interest, IUll also give a brief
introduction to Part II of the book, PHYSICAL WAVELETS, which is
research-level material:
Electromagnetic and acoustic wavelets
Applications to radar, sonar, and scattering.
If interested, please let me know so I have a rough idea of the size of
the class. Official registration is in January; call the grad school office
at (508) 934-2380. But if you cannot make a special trip, just show up in
class (after checking with me by phone or email) and IUll give you
some registration material that you can fill out to register by mail.
From: fournier@cloudy.geology.yale.edu
Subject: Answer: Use of Numerical Recipes software (WD 4.9 #10)
SUBROUTINE wt1tfa(a,n,aa,nmax,log_2n,wtstep) !Time/frequency analysis
!using Numerical Recipes wt1 by Press et al. (Cambridge, 2nd edition).
!This code by Aime' Fournier (fournier@cloudy.geology.yale.edu).
EXTERNAL wtstep !Requires n=2**log_2n.
REAL a(n) !Input signal becomes wavelet coefficient vector.
REAL aa(nmax,log_2n-1) !Output aa(.,j)=projection onto level j space.
CALL wt1(a,n,1,wtstep) !CALL pwtset before wt1tfa.
DO j=1,log_2n-1 !loop over timescales 2^j.
ncfcts=n/2**j !# coefficients at level j.
DO k=1,n !loop over time instants k.
IF ((ncfcts.LT.k).AND.(k.LE.2*ncfcts))THEN
aa(k,j)=a(k) !Keep only level j coefficients.
ELSE !mask coefficients corresponding other j levels:
aa(k,j)=0.
END IF
END DO
CALL wt1(aa(1,j),n,-1,wtstep) !Projection onto level j space.
END DO !The value aa(k,j) now corresponds to amplitude at time k
RETURN !and timescale 2^j.
END
From: Kurt TeKolste <tekolste.kurt@ist.vf.mmc.com>
Subject: Answer: Parametrization of Biorthogonal Wavelets (WD 4.10 #24)
A recent note from Thomas Sledge in this listserver prompted me to
look up some work that I did a few years ago. What I discovered at
that time was a way to generate all compactly-supported real valued
wavelets.
The generator is ridiculously easy to code, and I think that the
parameterization is suitable for doing the analysis necessary to
determine best wavelets for certain conditions.
The algorithms for finding dilation coefficients from my parameters
and vice versa follow.
Preliminary Definition
The construction hinges an idea which I call a "pivot." If {a_i | i
in Z} is a series, the "pivot of a through alpha at k," is {b_i | i in
Z} where b_i is determined by the formula: b_i = a_i * cos(alpha) +
(-1)^i * a_(2*k+1-i) * sin(alpha).
Dilation coeffiencets from parameters
Wavelets of "dimension" d (with dilation coefficents zero outside of
0, ..., 2d+1) are parameterized by d angles, alpha_1, alpha_2, ...,
alpha_d. To find the dilation coefficients, you proceed as follows.
Let alpha_0 = pi/2 - (alpha_1 + ... + alpha_d). Let a_0,0 =
cos(alpha_0/2), a_0,1 = sin(alpha_0/2). For j = 0, ..., d-1 let a_j+1
= {a_(j+1,i) | i in Z} be the series obtained from a_j = { a_(j,i) | i
in Z} by pivoting through alpha_(j+1) at j+1. Then a_d is a valid set
of dilation coefficients.
Segue
I do not think that this mapping is 1-1: there seem to be multiple
parameterizations for the same set of dilation coefficients. However,
it is onto, which we see next.
Parameters from Dilation Coefficients
This construction generates all the possible dilation coefficients.
This is demonstrated by showing how to find the alphas from the
dilation coefficients. We assume, without loss of generality, that it
is not the case that both a_0 and a_1 are zero and that a_i = 0 for i
< 0 and i > 2n+1. We find n angles as follows.
Let alpha_n = 2*arctan(a_2n+1/a_0) if a_0 <> 0 and alpha_n =
2*arctan(a_2n/a_1) if a_0 = 0. Pivot the original sequence by
-alpha_n at n. This yields a series with support in 0, ... , 2n-1.
Repeat for the shorter set of coefficients until you have alpha_0 =
2*arctan(a_1/a_0).
The resulting alpha_i will return the initial dilation coefficents via
the construction described above.
Thus, the zero dimensional wavelet is given by <cos(alpha_0/2),
sin(alpha_0/2)> where alpha_0 = pi/2.
The one dimensional wavelets are desribed by:
<cos(alpha0/2)cos(alpha1/2), sin(alpha0/2)cos(alpha1/2),
-sin(alpha0/2)sin(alpha1/2), cos(alpha0/2)sin(alpha1/2)> with alpha0 +
alpha1 = pi/2.
Since this is a bit tiresome, we use an obvious encoding so that the
0-wavelets are <c0, s0> the 1-wavelets are <c0*c1, s0*c1, -s0*s1,
c0*s1> Continuing, we have the 2-wavelets as <c0*c1*c2, s0*c1*c2,
-s0*s1*c2 - c0*s1*c2, c0*s1*c2 - s0*s1*s2, s0*c1*s2, c0*c1*s2> with
alpha0 + alpha1 + alpha2 = pi/2.
Colloquially, you get the next series by a) multiplying the current
series by cos(alpha), b) reversing the current series, alternating the
signs, multiplying by sin(alpha), and sliding it over two to the right
and then c) adding the two together. This is the basis of the
computer algorithm since it allows you to need to compute each cosine
and sine just once.
I wish I had an elegant Fourier analysis justification for this
construction, but I do not -- it is just a down and dirty discrete
algorithm as far as I have been able to work it.
I have built a "screensaver" which allows the user to wander around
the wavelet space of their choice up to dimension 150 (302
coefficients in the filters). It will also move regularly through the
collection of all wavelets or wander with randomly chosen
displacements.
Unfortunately, there is an error in the graphing routine which flips
the wavelet now and then. Since I no longer have the Pascal compiler
loaded it is too much trouble to fix it. If, with this caveat, you
would like a executable version of this MSDOS program sent via MIME,
please contact me via email.
Kurt TeKolste <tekolste.kurt@ist.vf.mmc.com>
Phone : 610-964-8791
From: Vladimir Bochev <VP007@BGEARN.BITNET>
Subject: Question: Looking for lapped filter banks
Hi,
I'll be extremely grateful is there is someone who can send
coefficients for 8, 12 or 16 band maximally decimated FIR filter
bank(parallel not hierarchical). I'm now currently using the 8 band
filter bank tabulated in Malvar's book, Signal Processing with lapped
transforms.
From: "Simon, Stephen (Steve)" <sis1@NIOBBS1.EM.CDC.GOV>
Subject: Question: Fractional Brownian Noise
I have a signal plus noise problem where the noise looks in many ways
like a fractional brownian process. I have lots of data with all
noise and no signal and in these cases, the variability in the wavelet
coefficients appear to follow a power law. I have been tinkering with
a thresholding rule that incorporates this power law. In simple
terms, I zero out the coefficients at the highest "frequency" if they
are within plus/minus K, at the next highest "frequency" if they are
within plus/minus K*2**alpha, at the next "frequency" if they are
within plus/minus K*2**(2*alpha), etc., where alpha is based by the
degree of self-similarity (self-affinity) of the noise. Traditional
thresholding rules preserve as part of the signal certain low
frequency patterns that are an artifact of the correlational nature of
the noise. So far, the power law seems to do a much better job of
separating signal from noise.
Before I go to a lot of trouble on this, is there any reference on the
use of wavelet decomposition on fractional brownian processes? I've
seen a few references on how wavelets "whiten" a process, but I want
to see results on the variability of the wavelet coefficients at each
"frequency" level.
Also, has anyone seen references on the use of a power law for
thresholding wavelet coefficients?
Steve Simon, sis1@niobbs1.em.cdc.gov
Standard disclaimer.
[Answer] [Answer] [Answer]
From: MEI@TRLVM.VNET.IBM.COM
Subject: Announce: Filter banks with more than 50 significant digits
Prof. Kazuo HATANO of Matsuyama, Japan developed a method to
accurately compute filter bank coefficients (50 and more significant
digits). And has complied tables for many well-known wavelets. For the
moment, only hard copies of the tables are available. I will submit
the tables to the major ftp site now under construction in Japan as
soon as the site is opened.
Mei Kobayashi
IBM Tokyo Research Laboratory
internet: mei@trlvm.vnet.ibm.com
gateway: mei@trl.ibm.co.jp
<webmaster@wavelet.org>