[Home]
[Back Issues]
[Search]
[Calendar]
[Vol 1]
[Vol 2]
[Vol 3]
[Vol 4]
[Vol 5]
[Vol 6]
[Vol 7]
[Vol 8]
The Wavelet Digest, Volume 8, Issue 1 (January 27, 1999)
Today's Editor: Wim Sweldens
Lucent Technologies, Bell Laboratories
wim@bell-labs.com
Today's Topics:
1: Book: Wavelets: An Analysis Tool
2: Book: Time-Frequency/Time/Scale Analysis.
3: Preprint: Wavelet bases in H(div) and H(curl)
4: Preprint: Three papers from Hiu Li on Turbulent Flow
5: Preprint: Translation invariance of wavelet
6: Preprint: N Dimensional Finite Wavelet Filters
7: Preprint: Decomposition for Recovery of Signal from Noisy Data
8: Preprint: Balanced Multiwavelet Bases Based on Symmetric FIR Filters
9: Thesis: Time-frequency and time-scale reassignment methods
10: Thesis: Wavelet Based Features for EMG Classification
11: Thesis: Wavelet-Galerkin Scheme of Electromagnetic Problems
12: Thesis: Beyond Traditional Transform Coding
13: Web: A wavelet tutorial after Stephane Mallat's book
14: Course: June 16-18, Braunschweig, Wavelet workshop by Gilbert Strang
15: Course: Mathematical Aspects of Image Processing
16: Job: Post-Doc Positions at the new national wavelet center
17: Contents: J. Approx. Th. - dec98/jan99
18: Answer: Generate all of the one dimensional filters (WD 7.11 #20)
19: Question: Wavelet Zero-Crossings for Image Compression
20: Question: Wavelets and financial engineering
21: Question: Wavelets on the interval
22: Question: Wavelet analysis for speech
23: Question: Wavelets for the analysis of planetary radio emissions
24: Question: Gaussian functions is a good basis function?
25: Question: Wavelet in Modal Anlysis
26: Question: Wavelets and data fusion
27: Question: I am looking for Commercial Smoothing software!
28: Question: Wavelets and ultrasound
29: Question: Evaluate Wavelet Functions in SPLus
30: Question: Wavelet Element Method
31: Question: Wavelets and classification
32: Question: Filters for music perception
Current number of members: 13250
E-mail:
General help: help@wavelet.org
Add yourself as a member: add@wavelet.org
Remove yourself as a member: remove@wavelet.org
Publish in the next WD: publish@wavelet.org
Receive publishing policy: policy@wavelet.org
Visit us on the Web at http://www.wavelet.org/
Calendar of events:
Feb 24-25: Wavelets, Royal Society Meeting, London
Feb 24-26: IEEE Information Theory Workshop, Santa Fe, NM
Mar 8-11: Course: Mathematical and Physical Wavelets, Boston, MA
Mar 17-19: CISS '99 (Information Sciences and Systems), Baltimore
Mar 22-26: II Symposium on Artificial Intelligence, Havana, Cuba
Apr 6-8 : SPIE Wavelet Applications VI, Orlando
Jun 14-19: Mathtools 99, St-Petersburg, Russia
Jun 14-16: Fractals in Engineering, Delft, The Netherlands
*Jun 16-28: Wavelet workshop by Gilbert Strang, Braunschweig
Jul 1-7: Curves and Surfaces, St-Malo, France
*Jul 6-16: Mathematical Aspects of Image Processing
Jul 18-23: SPIE Wavelet Applications VII, Denver, CO
Date: Mon, 21 Dec 1998 10:35:00 -0000
From: "PEACOCK, Tom" <peacockt@oup.co.uk>
Subject: #1 Book: Wavelets: An Analysis Tool
Wavelets: An Analysis Tool
M. Holschneider, C.N.R.S. Luminy, Centre de Physique Théorique,
Marseille, France
This popular book provides an explanation in elementary language of
the mathematical foundations of the theory of wavelet
analysis. Written for postgraduate students and researchers, this
timely introduction takes account of the continuous transform as well
as its discretized version, the ortho-normal basis. There are detailed
discussions of wavelet analysis of regular functions, tempered
distributions, and square integrable functions. The continuous wavelet
transform is deduced by means of various sampling theorems, and used
as a mathematical microscope to analyse fractals.
>From the reviews of the cloth edition: 'This book is a self-contained
and original text on the theory of wavelets ... Graduate students of
mathematics and physics will find it useful, as will interested
readers from other fields.' Physics World
'Holschneider has written a rich introduction to wavelet analysis that
has as its explicit aim to "fill a gap between the book by Ingrid
Daubechies ... and the more advanced books by Yves Meyer and Ronald
Coifman ... Hopefully this book will ease access to such post-graduate
literature". I think this is certainly being accomplished by this
work.' Dutch Mathematical Society Newsletter
CONTENTS
Introduction to wavelet analysis over IR
Discretizing and periodizing the half-plane
Multi-resolution analysis
Fractal analysis and wavelet transforms
Group theory as unifying language
Functional analysis and wavelets
Bibliography
Index
Oxford Mathematical Monographs
Paperback, 0-19-850521-3,448 pages, 27 halftones, 57 line drawings,
November 1998, GBP30.00
Order your copy by emailing book.orders@oup.co.uk or via our website
http://www1/orders/order.html
Date: Sun, 3 Jan 1999 10:48:30 +0100
From: flandrin@physique.ens-lyon.fr (Patrick Flandrin)
Subject: #2 Book: Time-Frequency/Time/Scale Analysis.
The English translation of "Temps-Fréquence", first published (in
French) in 1993 by Editions Hermès, Paris, is now available as Volume
10 in Chui's series "Wavelet analysis and its applications" :
P. Flandrin, "Time-Frequency/Time-Scale Analysis," Academic Press,
San Diego (CA), 1999.
ISBN 0-12-259870-9, 400 pages.
The book is a general and detailed survey of both time-frequency and
time-scale methods, with special emphasis on quadratic (Wigner- like)
transforms.
More info at http://www.apnet.com.
A 2nd edition of the French version is also available as
P. Flandrin, "Temps-Fréquence," Hermès, Paris, 1998.
ISBN 2-86601-700-5, 400 pages.
More info at http://www.editions-hermes.fr.
Patrick Flandrin
Laboratoire de Physique Ecole Normale Supérieure de Lyon
46 allee d'Italie 69364 Lyon Cedex 07 France
Tel.: +33 (0)4 72 72 81 60 Fax : +33 (0)4 72 72 80 80
Date: Tue, 22 Dec 1998 10:28:03 +0100
From: Karsten Urban <urban@dragon.ian.pv.cnr.it>
Subject: #3 Preprint: Wavelet bases in H(div) and H(curl)
The preprint
"Wavelet bases in H(div) and H(curl)"
by Karsten Urban
is available under
ftp://www.igpm.rwth-aachen.de/pub/urban/reports/hcurl.ps.Z
Abstract:
Some years ago, compactly supported divergence-free wavelets
have been constructed which also give rise to a stable
(biorthogonal) wavelet splitting of H(div). These bases have
successfully been used both in the analysis and numerical
treatment of the Stokes- and Navier-Stokes equations.
In this paper, we construct stable wavelet bases for the stream
function spaces H(curl). Moreover, curl-free vector wavelets are
constructed and analysed. The relationship between H(div) and
H(curl) are expressed in terms of these wavelets. We obtain
discrete (orthogonal) Hodge decompositions.
Our construction works independently of the space dimension,
but in terms of general assumptions on the underlying wavelet
systems in $L^2(\Omega)$ that are used as building blocks.
We give concrete examples of such bases for tensor product
and certain more general domains $\Omega\subset\er^n$.
As an application, we obtain wavelet multilevel preconditioners
in H(div) and H(curl).
Karsten Urban
present address:
Istituto di Analisi Numerica del C.N.R.
via Abbiategrasso, 209
I-27100 Pavia, Italy
urban@dragon.ian.pv.cnr.it
From: "Hui LI" <li@mech.kagoshima-u.ac.jp>
Subject: #4 Preprint: Three papers from Hiu Li on Turbulent Flow
(1)
Title:
Identification of Coherent Structure in Turbulent Shear Flow with Wavelet
Correlation Analysis
Author: Hui LI,
Ph.D., Assistant Professor
Department of Mechanical Engineering, Kagoshima University
1-21-40, Korimoto, Kagoshima 890-0065, JAPAN
Tel: +81-99-285-8252 Fax: +81-99-250-3181
E-mail: li@mech.kagoshima-u.ac.jp
Home Page: http://www.mech.kagoshima-u.ac.jp/~li/
Status:
ASME Journal of Fluids Engineering, Vol.120 (1998), No.4, pp.778-785
Abstract:
In order to identify coherent structure of turbulent shear flow, a
new combination of familiar techniques of signal processing, called
wavelet correlation analysis, is developed based on the wavelet
transform. The wavelet correlation analysis provides the unique
capability for decomposing the correlation of arbitrary signals over a
two-dimensional time delay-period plane. By analyzing two
superposition functions implicating several pure frequencies, the
correlation of periodic oscillations at several frequencies can well
be separated and observed clearly. Coherent structures in the
intermediate region of a plane turbulent jet are investigated using
the wavelet correlation method. It is shown that the wavelet
correlation analysis can extract the most essential scales governing
features of eddy motions. The coherent structure information and
apparent flapping behaviors are clearly revealed over a
two-dimensional time-period plane.
This preprint can be downloaded at
http://www.mech.kagoshima-u.ac.jp/~li/
or
http://www.asme.org/pubs/journals/fluideng/JFE.html (FREE until June 30,
1999)
(2)
Title:
Wavelet Statistical Analysis of the near Field Flow Structure in a
Turbulent Jet
Author: Hui LI,
Ph.D., Assistant Professor
Department of Mechanical Engineering, Kagoshima University
1-21-40, Korimoto, Kagoshima 890-0065, JAPAN
Tel: +81-99-285-8252 Fax: +81-99-250-3181
E-mail: li@mech.kagoshima-u.ac.jp
Home Page: http://www.mech.kagoshima-u.ac.jp/~li/
Status:
Transactions of the Japan Society for Aeronautical and Space
Sciences, Vol.41 (1998), No.133, pp.125-132
Abstract:
In present study, the definitions of the wavelet-based statistical
quantities on turbulence analysis were first developed, in order to
study the characteristics of statistics in frequency space. Then, the
wavelet statistical quantities were used to analyzing eddy motions and
the structure of the energy transfer over a two-dimensional
frequency-physics plane in a plane jet. In the shear layer, behaviors
of Reynolds stress reversals were easily detected at several
frequencies at x/b0=2, 6 and 10. This phenomenon indicated that
fluctuation energy is transferred from high frequency component to low
frequency component. The frequency of wavelet Reynolds stress reversal
decreases toward downstream and the interaction between positive and
negative production in frequency space leads to the formation of the
coherent structures.
more information about this work is available from my website, at
http://www.mech.kagoshima-u.ac.jp/~li/
(3)
Title:
Effect of Different Orthogonal Wavelet Basis on Multiresolution Image
Analysis of a Turbulent Flow
Authors: Hui LI
M. TAKEI
M. OCHI
Y. SAITO
K. HORII
Status:
CD-ROM Proceedings of International Conference on Optical Technology
and Image Processing
in Fluid, Thermal and Combustion Flow, Japan (1998), No.AB015, pp.1-11.
Abstract:
This study develops a new application of two-dimensional
orthogonal wavelets to turbulence, in order to reveal the
multiresolution turbulent structures and to extract the coherent
structure. The digital imaging slice photographs of turbulent jet in
a far field with Re=4500 and 9000 are respectively decomposed into six
image components with different broader scales by the wavelet
multiresolution analysis. These image components provide further
evident of the multiscale structures in turbulent jet. It was found
that the edges of the vortices at different resolutions or scales and
the coherent structure might be easily extracted. In this problem,
three types of image structure, namely the large-scale structure near
the center of jet, coherent structure in the shear layer and the
small-scale vortices in the whole flow field, which dominate the
turbulent structure, were clearly displayed. With increasing Reynolds
number, the large-scale structure becomes weaker, and the scale of
active vortices and the scale of coherent structure decreases.
This preprint can be downloaded at
http://www.mech.kagoshima-u.ac.jp/~li/
From:
Date: Sun, 10 Jan 1999 15:28:49 +0800
From: "qiaowang" <qiaowang@seu.edu.cn>
Subject: #5 Preprint: Translation invariance of wavelet
Preprint: Direct and inverse problems for translation invariance of wavelet
decomposition, by Qiao WANG and Lenan WU
ABSTRACT: Most of the wavelet bases haven't the shiftability, thus
there exist a pair of fundamental problems: The direct one is to
estimate the wavelet expansion of translating signal from that of the
translated signal. The inverse one is to estimate the shift quantity
from expansions of two signals with the same waveform. In this paper,
we develope our metric functional method and furthermore introduce a
group of basic interpolation functions. Exactly saying, they are $$
r_n(\phi)={1\over 2\pi}{\int \left | \widehat\psi(\xi) \widehat
\psi(2^(-n)\xi) \right |^2 d\xi } $$ and $$\Delta_{n,s}=\int
2^{n/2}\psi(2^nx-s)\overline {\slide \psi(x)} dx $$
respectively. Using the functional defined above, we found that some
commonly used wavelet bases such as Chui-Wang's semi-orthonormal
wavelets, Battle-Lemarie's wavelets etc., are nearly shiftable. The
direct and inverse problems are discussed in details for these nearly
shiftable wavelets, and a translation invariant (orthonormal or
semiorthonormal) wavelet decomposition algorithm is proposed. The
error analysis are built in a statistical-like way based on functional
$r_n(\psi)$ and/or $\| \Delta _{n,s} \|$, and the equivalence between
$r_n(\psi)$ and/or $\| \Delta _{n,s} \|^2$ is proved.
(This work is supported by Post-Doctor Research Foundation of China and NSF of China. )
If you need, please send us an e-mail to qiaowang@seu.edu.cn
Dr. Qiao WANG
Graphics & Robotics Laboratory
Dept of Radio Engineering, Southeast University
Nanjing, 210096 P.R.China
Date: Fri, 15 Jan 1999 22:23:11 +0800
From: "Peng Silong" <silence.peng@263.net>
Subject: #6 Preprint: N Dimensional Finite Wavelet Filters
Abstract
In this paper, a large class of n-dimensional orthogonal wavelet
filters (lowpass and highpass) are presented in explicit
expression. In particularly in one dimension case, this expression is
complete, that means, all one dimensional orthogonal wavelet filters
are included. We also characterize the filters with linear phase in
this case. To ensure these filters are reasonal, that is, they can
generate wavelet functions, the necessary condition of othogonal are
also proposed. In addition, relations between two filters are
discussed. Some examples are also given, including nonsparable filters
with linear phase.
Peng Silong
NADEC, Institute of Automation,
Beijing, 100080, P.R.China
email:silence.peng@263.net
or
pengsl@263.net
Date: Tue, 19 Jan 1999 08:33:33 +0800
From: Hsin-Cheng Huang <hchuang@stat.sinica.edu.tw>
Subject: #7 Preprint: Decomposition for Recovery of Signal from Noisy Data
Title: Deterministic/Stochastic Wavelet Decomposition for Recovery of
Signal from Noisy Data
Authors:
Hsin-Cheng Huang, Institute of Statistical Science, Academia Sinica,
Taipei 115, Taiwan
hchuang@stat.sinica.edu.tw
Noel Cressie, Department of Statistics, The Ohio State University,
Columbus, OH 43210-1247, U.S.A.
ncressie@stat.ohio-state.edu
Abstract:
In a series of recent papers on nonparametric regression, D. Donoho
and I. Johnstone developed wavelet shrinkage methods for recovering
unknown piecewise-smooth deterministic signals from noisy
data. Wavelet shrinkage based on the Bayesian approach involves
specifying a prior distribution on the wavelet coefficients, which is
usually assumed to have a distribution with zero mean. There is no a
priori reason why all prior means should be zero; indeed, one can
imagine certain types of signals where this is not a good choice of
model. In this paper, we take an empirical Bayes approach where we
propose an estimator for the prior mean that is "plugged into'' the
Bayesian shrinkage formulas. Another way we are more general than
previous work is that we assume the underlying signal is composed of a
piecewise-smooth deterministic part plus a zero-mean stochastic part;
that is, the signal may contain a reasonably large number of nonzero
wavelet coefficients. Our goal is to predict this signal from noisy
data. We also develop a new estimator for the noise variance based on
a geostatistical method that considers the behavior of the variogram
near the origin. Simulation studies show that our method
(DecompShrink) outperforms the well known VisuShrink and SureShrink
methods for recovering a wide variety of signals. Moreover, it is
insensitive to the choice of the lowest-scale cut-off parameter, which
is typically not the case for other wavelet shrinkage methods.
The paper is available at:
http://wagner.stat.sinica.edu.tw/~hchuang/paper.html
Date: Wed, 20 Jan 1999 17:32:36 -0500 (EST)
From: Ivan Selesnick <selesi@taco.poly.edu>
Subject: #8 Preprint: Balanced Multiwavelet Bases Based on Symmetric FIR Filters
Balanced Multiwavelet Bases Based on Symmetric FIR Filters
This paper describes a basic difference between multiwavelets and
scalar wavelets that explains, without using zero moment properties,
why certain complications arise in the implementation of discrete
multiwavelet transforms. Assuming one wishes to avoid the use of
prefilters in implementing the discrete multiwavelet transform, it is
suggested that the behavior of the iterated filter bank associated
with a multiwavelet basis of multiplicity r is more fully revealed by
an expanded set of r^2 scaling functions \phi_{i,j}.
This paper also introduces new K-balanced orthogonal multiwavelet
bases based on symmetric FIR filters. The nonlinear design equations
arising in this work are solved using Groebner bases. The K-balanced
multiwavelet bases based on even-length symmetric FIR filters are
shown to be particularly well behaved, as illustrated by special
relations they satisfy and by the smooth examples constructed.
Ivan Selesnick
Electrical Engineering
Polytechnic University
6 Metrotech Center
Brooklyn, NY 11201, USA
selesi@taco.poly.edu
http://taco.poly.edu/selesi/
Date: Mon, 21 Dec 1998 15:33:29 +0100 (MET)
From: Chassande-Mottin <eric@aei-potsdam.mpg.de>
Subject: #9 Thesis: Time-frequency and time-scale reassignment methods
PhD Thesis: Reassignment methods in the time-frequency plane for the
analysis and processing of non stationary signals by Eric
Chassande-Mottin, Laboratoire de Physique de l'ENS Lyon
Comittee:
Francois Auger Patrick Duvaut
Richard G. Baraniuk Patrick Flandrin -- Advisor
Christian Doncarli Bruno Torresani
Abstract:
We are interested in the so-called reassignment method. This method is
devoted to improve the readability of time-frequency or time-scale
representations through the action of a (reassignment) vector field
which moves the values of the time-frequency distribution. Although
the principle of reassignment is well known, this is not the case with
the consequences of its use. The first goal of this thesis is to give
some elements for answering this question. For example, we perform a
statistical description of the spectrogram reassignment vector field
for a noisy signal. In a second step, by taking this information into
account, we put forward improvements to the original reassignment
method : we propose a supervision technique which decides in each
time-frequency point whether it is convenient to reassign or not.
Finally, we show that the reassignment method can be useful for tasks
beyond analysis and can be part of a signal processing chain: by using
a precise geometrical description of the reassignment vector field, we
introduce a new reassignment method, referred to as differential
reassignment, with which it becomes possible to tile the
time-frequency plane in regions. We associate each of these regions to
a signal component. In another context but with the same objective of
using reassignment to do signal processing tasks, we show that the
reassigned time-frequency distributions can be useful for chirp
detection problems, such as the gravitational wave one.
This thesis is written in French.
Postcript and PDF files of the manuscript can be obtained at the
following address:
http://www.physique.ens-lyon.fr/ts/these/ecm98.html
The Matlab source codes for the computation of all the figures of this
thesis are also available at this address.
Date: Wed, 30 Dec 1998 10:03:00 -0700
From: "Kevin Englehart" <kengleha@unb.ca>
Subject: #10 Thesis: Wavelet Based Features for EMG Classification
Dear Colleagues,
My Ph.D. thesis was completed in October of 1998 in the department of =
Electrical and Computer Engineering at the University of New Brunswick, =
Canada.
Title: Signal Representation for Classification of the Transient =
Myoelectric Signal
Supervisors: Philip Parker, Maryhelen Stevenson
This work demonstrates the application of wavelet and wavelet packet =
based features for EMG pattern classification. The abstract, and the =
entire manuscript is available at
http://sern.ucalgary.ca/~kengleha/thesis/phd.htm
I would like to acknowledge the insights gained via the readership of =
the Wavelet Digest, especially that of Fred Warner (Yale University) and =
Naoki Saito (UC Davis).
Best Regards,
Kevin Englehart
Kevin Englehart, Ph.D., P.Eng.
Assistant Professor, Electrical and Computer Eng.
University of New Brunswick
P.O. Box 4400, Fredericton, N.B., Canada, E3B 5A3
Tel: +1 506-453-4561 Fax: +1 506-453-3589
http://sern.ucalgary.ca/~kengleha
Date: Thu, 21 Jan 1999 12:35:32 KST
From: <nopride@hanmail.net>
Subject: #11 Thesis: Wavelet-Galerkin Scheme of Electromagnetic Problems
Dear wavelet friends and colleagues:
In Jan of 99 I completed my Master thesis
at KwangWoon University, Seoul, South Korea.
Title: Wavelet-Galerkin Scheme of Inhomogeneous Electromagnetic Problems
Using the Time-Dependent Maxwell's Equations
Author: Young Wook Cheong (e-mail: nopride@hanmail.net)
Advisor: Keuk Hwan Ra
Abstract:
A wavelet-Galerkin scheme based on the time-dependent Maxwell's
equations is presented. Daubechies' wavelet with two vanishing
wavelet moments is expanded for basis function in spatial domain
and Yee's leap-frog approach is applied. The shifted interpolation
property of Daubechies' wavelet family leads to the simplified
formulations for inhomogeneous media without the additional
matrices for the integral or material operator. The stability
condition is formulated. The dispersion characteristics are
analyzed and compared with those of finite difference time domain and
multiresolution time domain methods. The analyses show the excellent
trade-off between the regularity and the support width of the basis
function. Although the basis function has only two vanishing wavelet
moments, it is enough to provide negligible dispersive error in the
numerical analysis and its compact support enables only several
involved terms per nodes. The storage effectiveness, execution time
reduction and accuracy of this scheme are demonstrated by calculating
the resonant frequencies of the homogeneous and inhomogeneous
cavities.
If you want to obtain an abstract, contents, and the whole thesis in PDF format,
then send me E-mail (nopride@hanmail.net).
Date: Tue, 26 Jan 1999 22:40:40 -0500
From: "Vivek Goyal" <vivek@research.bell-labs.com>
Subject: #12 Thesis: Beyond Traditional Transform Coding
PhD Thesis: Beyond Traditional Transform Coding
Author: Vivek Goyal
Date: December 1998
Institution: University of California, Berkeley
Advisor: Martin Vetterli
Abridged Abstract:
Transform coding is the most successful and pervasive technique
for lossy compression of audio, images, and video. In conventional
transform coding, the original signal is mapped to an intermediary
by a linear transform; the final compressed form is produced by
scalar quantization of the intermediary and entropy coding. This
thesis extends the traditional theory toward several goals: improved
compression of nonstationary and non-Gaussian signals; robust joint
source-channel coding for erasure channels; and computational
complexity reduction or optimization.
First the use of frames, overcomplete sets of vectors, is
explored. Linear transforms based on frames give representations
with robustness to random additive noise and quantization.
Nonlinear, signal-adaptive representations can be produced with
frames using matching pursuit. This algorithm exhibits good
compression performance at low rates. Optimal reconstruction is
described for both linear and nonlinear frame-based representations.
Within the conventional setting of basis representations, the
Karhunen- Loeve transform (KLT) gives optimal compression of
Gaussian sources. Its dependence on the probability density of the
source, which is generally unknown, limits its application.
Including a description of an estimated KLT in the data creates
significant overhead. In contrast, this thesis introduces a method
for on-line universal transform coding which utilizes backward
adaptation and has asymptotically negligible overhead. Also
developed are techniques inspired by adaptive FIR Wiener filtering.
Transform coding is normally used solely for source coding.
This thesis introduces two joint source-channel coding methods for
erasure channels, presented in the context of multiple description
coding. These give graceful performance degradation in the face of
random loss of transform coefficients. One technique produces
correlated transform coefficients so that lost coefficients can be
estimated. The rate allocated to channel coding is continuously
adjustable with fixed block length. The second technique--
quantized frame expansion--is similar to scalar quantization
followed by a block channel code, except that quantization and
introduction of redundancy are swapped. In certain circumstances
the end-to-end performance significantly exceeds that of a system
with separate source and channel coding.
Finally, joint optimization of computational complexity and
rate-distortion performance is considered. Sample analyses compare
unstructured and harmonic transforms, and show that JPEG encoding
complexity can be reduced with little loss in performance.
The complete abstract is available on-line at
/cm/ms/who/vivek/Thesis/
Instructions for downloading the entire thesis appear at the same address.
Your comments are always welcome. Thank you.
Vivek K Goyal
Mathematics of Communications Research, Bell Labs
Office: 600 Mountain Avenue +1 908 582 6484
Room 2C-178 fax: 582 3340
Murray Hill, NJ 07974
E-mail: v.goyal@ieee.org
URL: /who/vivek/
Date: Fri, 22 Jan 1999 11:59:28 +0100
From: Francois Chaplais <chaplais@cas.ensmp.fr>
Subject: #13 Web: A wavelet tutorial after Stephane Mallat's book
A Web presentation of Stephane Mallat's Book "A Wavelet Tour
of Signal Processing" can be browsed at the following address
http://cas.ensmp.fr/~chaplais/Wavetour_presentation/Wavetour_presentation_US.html
The presentation uses the implicit links between subjects that can be
found in the paper book. It uses frames to make easier the comparision
between similar methods (for instance, windowed Fourier
v.s. wavelets). It gives a detailed insight on the contents of most of
the book (except for the last chapters). By the way, a revised second
US edition should be available during April (after S. Mallat's words).
A French version of the presentation is located at
http://cas.ensmp.fr/~chaplais/Wavetour_presentation/Wavetour_presentation_fr.html
A starting point for both languages is
http://cas.ensmp.fr/~chaplais/Wavetour_presentation/
This Web version was written last July after I had the pleasure of
translating S. Mallat's book into French. The French edition should be
out in September and include some modified material. Comments are
welcome. Enjoy!
F.Chaplais
Centre Automatique et Systemes
35 rue Saint Honore 77305 Fontainebleau France
e-mail: chaplais@cas.ensmp.fr
http://cas.ensmp.fr/~chaplais/
fax: (33) 1 64 69 47 01
Date: Mon, 18 Jan 1999 10:31:00 +0100
From: Andreas Keese <a.keese@tu-bs.de>
Subject: #14 Course: June 16-18, Braunschweig, Wavelet workshop by Gilbert Strang
FIRST ANNOUNCEMENT
WORKSHOP COURSE ON
WAVELETS AND FILTER BANKS
taught by Gilbert Strang(MIT)
with selected applications presented and practical computer training
directed by other researchers
Wednesday-Thursday-Friday
June 16-18, 1999
at the Technical University of Braunschweig, Germany
TOPICS We will aim for the right balance of theory and applications. The
workshop course gives an overall perspective of the field, following the
program of six successful workshops by Professors Strang and Nguyen in the
US. The program will develop these four key areas in detail:
1. Analysis
Multirate Signal Processing: Filtering, Decimation, Polyphase
Perfect Reconstruction and Aliasing Removal
Matrix Analysis: Toeplitz Matrices and Fast Algorithms
Wavelet Transform: Pyramid and Cascade Algorithms
Daubechies Wavelets, Orthogonal and Biorthogonal Wavelets
Smoothness, Approximation, Boundary Filters and Wavelets
Time-Frequency and Time-Scale Analysis
2. Design Methods
Spectral Factorization
Cosine-Modulated Filter Banks
Design by Constrained Least Squares
Lattice Structure
Ladder Structure (Lifting)
3. Applications (from Lecturers and Participants)
Audio and Image Compression, Quantization Effects
Transient Detection and Edge Detection
Digital Communication and Multicarrier Modulation
Transmultiplexers
Text-Image Compression: Lossy and Lossless
Medical Imaging and Scientific Visualization
Image Compression / Image Segmentation / Image Enhancement
Video Compression
Analysis of Electrodynamic Problems
4. Hands-on Experience with Software
MATLAB Wavelet Toolbox
Software for Compression
The goal of the Workshop is to be as useful as possible to all
participants. PLEASE bring to the workshop any application areas that you
would like to hear discussed. This is a very active workshop !
TEXTBOOK Participants will receive the textbook
WAVELETS AND FILTER BANKS by Strang and Nguyen
Wellesley-Cambridge Press, ISBN 0-9614088-7-1
This text is already in class use in many Electrical Engineering and
Mathematics departments. It was chosen to accompany MATLAB's Wavelet Toolbox,
which will be the simulation software at the Braunschweig Wavelet Workshop.
FURTHER INFORMATION AND REGISTRATION It is good to register early since only a
limited number of places is available. Further information on tuition, social
events, accomodation, travel directions and workshop registration is available
online on the web page
http://www.tu-bs.de/institute/WiR/wavelet99
or by email to wavelet99@tu-bs.de or by mail/phone/fax to
Technische Universit"at Braunschweig
Zentralstelle f"ur Weiterbildung
Postfach 33 29
D-38023 Braunschweig
phone +49/531/391-4211
fax +49/531/391-4215
Date: Fri, 15 Jan 1999 10:41:34 +0100
From: Consol Roca <crm@bianya.crm.es>
Subject: #15 Course: Mathematical Aspects of Image Processing
ADVANCED COURSE ANNOUNCEMENT
Advanced course on mathematical aspects of image processing.
July 6 - 16, 1999 at the Centre de Recerca Matemàtica and
Centre de Visió per Computador,
Campus of the Universitat Autònoma de Barcelona,
08193 Bellaterra (Spain)
PROGRAMME
The activity will consist in three courses:
1. Speakers: Dr. V. Caselles (U. Illes Balears) and
Dr. J.M. Morel (É.N.S., Cachan)
Title: "Mathematical models in image processing"
Summary: The role of partial differential equations in image
processing will be explained by the modelling of some basic
problems: image contrast, invariant smoothing, image interpolation
and classification of non linear filters. Applications will be
given to image intersection, denoising and disocclusion, shape
analysis, image interpolation and deblurring (18 hours).
2. Speaker:Dr. J. Serra (École des Mines, Paris)
Title: "Mathematical morphology"
Summary: Three aspects of mathematical morphology will be
presented. The first one, morphological filtering, involves
idempotent operators on complete lattices and connected operators.
The second aspect deals with segmentation based on the watershed
approach. Both aspects are then associated to multimedia
applications (indexation, tracking, compression). (12 hours).
3. Speakers: Dr. S. Mallat (É. Polytéchnique, Palaiseau) and
Dr. Y. Meyer (É.N.S., Cachan)
Title: "Signal processing with wavelets"
Summary: The wavelet transform, the wavelet orthonormal bases
associated to multiresolution approximations and their relation
with filter banks will be presented. Applications will be given to
edge detection, image compression and its relation with
approximation theory, linear and non-linear denoising and the
theory of minimax estimation. (18 hours).
SCIENTIFIC COORDINATORS
Joaquim Bruna and
Jordi Saludes
THIS ACTIVITY IS SUPPORTED BY THE EUROPEAN COMMISSION UNDER CONTRACT
ERBFMMACT980428 OF THE TMR PROGRAMM
REGISTRATION, ACCOMMODATION AND GRANTS
To obtain all this information, e-mail crm@crm.es or visit the web site
http://crm.es/info/acmaip/acmaip.html. Deadline for registration is May
15 (April 6 for those applying for grants)
Joaquim Bruna
Date: Mon, 25 Jan 1999 11:57:39 -0600 (CST)
From: Amos Ron <amos@cs.wisc.edu>
Subject: #16 Job: Post-Doc Positions at the new national wavelet center
Topic: post doc positions at the new national wavelet center.
The newly founded Wavelet Center for Ideal Data Representation is a
consortium of seven institutions that is aimed at developing new
wavelet/multiscale theories and technologies based on the principle of
REDUNDANT REPRESENTATIONS. Its secondary goal is to serve as a
national center for the entire development in the area of
multiscale/wavelet methods. The current participating institutions are
AT&T-Labs, Caltech, RPI, Princeton, Stanford, U. South Carolina and
U. Wisconsin. Current affiliates and collaborators of the center are
from Lucent Tech., MIT, U. Tel-Aviv, and U. Vienna. The headquarters
of the center is in Madison, Wisconsin. You may find more information
about the IDR center at http://www.waveletidr.org (Note that the site
is still under construction, so visit us soon again)
We currently offer three or more Post-Doctoral Positions at the
participating institutions (in () are the center's members and
affiliates who reside at the corresponding institutions):
One or more positions at PRINCETON, partly joint with AT&T.
(I. Daubechies, M. Orchard, Y. Kevrekidis, W. Willinger, A. Gilbert)
One position jointly at STANFORD and U. WISCONSIN.
(D. Donoho, E. Candes, C. de Boor, A. Ron)
One position at U. SOUTH CAROLINA.
(R. DeVore, R. Sharpley)
Each position is offered for an initial period of one year, beginning
Fall 1999, and may be renewed for an additional 1-2 years. The
positions offer a unique opportunity for recent Ph.D. to participate
in challenging problems in wavelet-related research. Successful
applicants will work with a member of the center at the hiring
institution. They will also be required to conduct portions of their
research activity at the other institutions of the center.
We invite applications from outstanding candidates. Applicants should
have a Ph.D. in a related field and have demonstrated strong research
potential. To receive full consideration, the application forms must
reach us by March 15, 1999. Applications will be considered until all
the positions are staffed.
Forms and other application materials can be found at
http://www.waveletidr.org
Applications in response to this solicitation MUST be submitted
electronically to the headquarters in Wisconsin through the
application Submission Form at the Web site. (Applications by regular
mail sent to the headquarters will not be considered or acknowledged,
and the submission materials will not be returned).
(-) Amos Ron
Coordinator
The Wavelet IDR Center
Date: Mon, 21 Dec 1998 00:21:06 -0500 (EST)
From: Thomas Hogan <hogan@math.ohio-state.edu>
Subject: #17 Contents: J. Approx. Th. - dec98/jan99
Table of Contents: J. Approx. Theory, Volume 95, Number 3, December 1998
A. L. Lukashov
On Chebyshev-Markov rational functions over several intervals
333--352
H. Hudzik and W. Kurc
Monotonicity properties of Musielak-Orlicz spaces and dominated best
approximation in Banach lattices
353--368
Xiaoming Zeng and Alain Piriou
On the rate of convergence of two Bernstein-B\'ezier type operators for
bounded variation functions
369--387
P. Deift, T. Kriecherbauer, and K. T.-R. McLaughlin
New results on the equilibrium measure for logarithmic potentials in the
presence of an external field
388--475
Roy Jones, Xin Li, R. N. Mohapatra, and R. S. Rodriguez
On the Bernstein inequality for rational functions with a prescribed
zero
476--496
Note
Horst Alzer
A refinement of Carleman's inequality
497--499
Author index for volume 95
500
Table of Contents: J. Approx. Theory, Volume 96, Number 1, January 1999
Leonid Golinskii, Paul Nevai, Ferenc Pint\'er, and Walter Van Assche
Perturbation of orthogonal polynomials on an arc of the unit circle, II
1--33
Vladimir Logvinenko and Alexander Russakovskii
Quasi-sampling sets for analytic functions in a cone
33--45
Michael S. Floater
Total positivity and convexity preservation
46--66
H. N. Mhaskar and J. Prestin
Bounded quasi-interpolatory polynomial operators
67--85
Agnieszka Lipieta
Cominimal projections in $l^n_\infty$
86--100
Guang Shi Ying
On Tur\'an quadrature formulas for the Chebyshev weight
101--110
Paul Butzer and Fran\c{c}ois Jongmans
P. L. Chebyshev (1821-1894), a guide to his life and work
111--138
Daniel Wulbert
Interpolation at a few points
139--148
Leszek Skrzypczak
Heat and harmonic extensions for function spaces of Hardy-Sobolev-Besov
type on symmetric spaces and Lie groups
149--170
Date: Mon, 18 Jan 1999 01:16:31 +0800
From: "Peng Silong" <silence.peng@263.net>
Subject: #18 Answer: Generate all of the one dimensional filters
(WD 7.11 #20)
>I am new to Wavelets.
>I am working with some signals composed of a number (1 - 100, it
>varies) of decaying sine waves, plus noise.
>I have started using wavelets to process these signals to reduce the
>noise, using the codes published in "Numerical Recipes in C" (second
>edition) Chapter 13.10.
>Questions
>(1) Are the coefficients of the Daubechies wavelets available to full
>double precision? If so, where please?
>(3) Are these in fact the best wavelets to use for my problem?
>(4) If not, please recommend me to something. For my application,
>code should preferably be in C or C++. I can translate FORTRAN.
Dear Dr John P. Fletcher,
I would like to answer your above questions.
(1)(4), I have writen a program based on my new paper on this essue to
compute all the one dimensional othogonal wavelet filters(not only the
famous Daubechies wavelets given in her book, but also many others). The
program is written in Matlab5.1, I wish you can read it and translate it to
C/C++.
(3), In fact, the filters in Daubechies book are not always the best, for
some applications, some other filters do better, it depend on your
application. You can use my program to test which is the best, if so, I hope
to hear from you about your application.
Best wishes
Peng Silong
Date: Sun, 20 Dec 1998 13:30:52 +0530 (IST)
From: "S. Sriram" <ssram@protocol.ece.iisc.ernet.in>
Subject: #19 Question: Wavelet Zero-Crossings for Image Compression
Hi, I'm interested in looking at Wavelet Zero-crossings for Image
Compression for my masters project. Can any of you please give me
some pointers to materials on this topic or send copies of relevant
papers to the address below.
Thanking you in advance
Sriram
S-80 IISc Hostel
IISc Bangalore
Karnataka 560 012
India
ssram@protocol.ece.iisc.ernet.in
Date: Tue, 22 Dec 1998 20:35:53 +0800
From: songbo@sklec.ecnu.edu.cn (Song Bo)
Subject: #20 Question: Wavelets and financial engineering
Hi,
I am a teacher in the college of management ,Shanghai JiaoTong =
University.As a new comer of the wavelets field, I would really =
appreciate if any of you can give me any reference regarding the =
application of wavelets to financial engineering. Thanks.
Liu Xin
Email:songbo@sklec.ecnu.edu.cn
Date: Wed, 23 Dec 1998 16:49:45 +0100
From: Armelle DAULLE <daulle@labs.polycnrs-gre.fr>
Subject: #21 Question: Wavelets on the interval
Some time ago, I asked about constructing Cohen-Daubechies-Vial
wavelets on the interval, and I want to thank all of you who helped me
some way. Now that I am done with constructing those edge wavelets,
my problem stands in the way to use the basis on the interval when
boundary conditions are not homogenous. More precisely, what's the
use of having N wavelets on each edge, when only just one would be
sufficient to meet boundary requirements? Is there any of you having
already implemented and used this particular basis (especially within
the frame of Galerkin method with non-homogenous boundary conditions)
who could supply me with some clue ?
Thanks in advance
Armelle DAULLE
Date: Thu, 24 Dec 1998 07:32:55 +0800
From: "KC Chan" <kcchan2@ee.cuhk.edu.hk>
Subject: #22 Question: Wavelet analysis for speech
Dear ALL
I am doing a proposal about speech recognition. Does anyone who knows
= where I can get some information about how wavelet can be used in
the = analysis of speech signal? I did come across some information
in the = internet, it says wavelet doesn't suit for speech signal.
Regards
KC
Date: Wed, 6 Jan 1999 10:19:30 +0000
From: "Michael Leitner" <mil@bimgs5.kfunigraz.ac.at>
Subject: #23 Question: Wavelets for the analysis of planetary radio emissions
Dear Fellows,
I am working on a project connected to the analysis of planetary radio
emissions. I want to collect experience about the 1D Wavelet
transformation of received time series of planetary radio emissions.
Has anyone of you worked with the wavelet transformation of
nonartifical signals? In detail I am interested in the advatages you
have by using the wavelet transformation compared to the Fourier
transformation. Finally if you are using wavelets for the analysis of
maybe completely different signals, like accoustical waves , or other
types please contact me. So we could exchange our experience.
Michael
DI Michael Leitner
Space Research Institute
Austrian Academy of Sciences
Halbaerthgasse 1
8010 GRAZ
AUSTRIA
Tel.: 0043 316 380 5264
Fax.: 0043 316 380 9825
email: michael.leitner@oeaw.ac.at
Date: Thu, 07 Jan 1999 13:09:48 +1100
From: Dennis Lee <dennis.lee@eng.monash.edu.au>
Subject: #24 Question: Gaussian functions is a good basis function?
Dear Wavelet experts
My question is that how to select a good basis function for
vibration transient wavelet analysis. Gaussian functions may possess
good approximation properties. If we select the scaling function of a
Gaussian, then what should we select for the wavelet function in order
to satisfy the wavelet basis admissibility conditions? I shall be
grateful if any experts may give me some insights? Your emails will
be submitted as answers in the next wavelet digest. Thanks!
regards
Dennis Lee
Date: Thu, 07 Jan 1999 17:14:58 -0200
From: Marcelo Braga dos Santos <mbsantos@mecanica.ufu.br>
Subject: #25 Question: Wavelet in Modal Anlysis
I am an mechanical engineer and my master degree are about modal
anlysis. I am interest in work with wavelets for parameter estimation
of mechanical system, natural frequency, damping factor and
eigenvectors. If anyone have works, references or sugestions, please
send for me.
Thanks
Marcelo Braga dos Santos
Federal University of Uberlandia - Brazil
Date: Fri, 8 Jan 1999 15:31:02 -0300 (GMT)
From: ioga@syfed.refer.mg
Subject: #26 Question: Wavelets and data fusion
Dear Wavelet Digest experts,
I have just one year of experiences about wavelet. My topic of
research is about data fusion problem in remote sensing using
wavelet and multiresolution analysis. I need any materials
about this field (references, documentation, software,...). I
already thank you for your help and kindness.
Best regards for everybody
Solofoarisoa RAKOTONIAINA
Institut & Observatoire Geophysique
d'Antananarivo (I.O.G.A)
B.P 3843
101, Antananarivo
MADAGASCAR
Tel/Fax: (261) 20-22 253 53
e-mail: ioga@syfed.refer.mg
Date: Fri, 08 Jan 1999 07:49:25 PST
From: "rosemarie barnett" <rosemariebarnett@hotmail.com>
Subject: #27 Question: I am looking for Commercial Smoothing software!
Hi, I am looking for commercial software that smooths data using
wavelets. Also does the wavelet smoothing algorithm require the
data to be equally spaced?
[Answer WD 8.2 #14]
Date: Wed, 23 Dec 1998 15:23:13 -0500
From: VccAso <vccaso@cidet.icmf.inf.cu>
Subject: #28 Question: Wavelets and ultrasound
Hi, I am an electrical engineering with specialization in Ultrasound
signal Processing. Off late I have developed a keen interest in the
area of Wavelets. I would be glad if anybody can help me in the
use of wavelet in reducing noise in ultrasonic signals in
concrete.Any sort of help is welcome.
With the best regards
Victor Castro
E-mail: vcc@uxdea4.iimas.unam.mx, vccaso@cidet.icmf.inf.cu
Date: Tue, 19 Jan 1999 10:05:12 +1100
From: ShinnJu-Lin <shinn-juh.lin@uts.edu.au>
Subject: #29 Question: Evaluate Wavelet Functions in SPLus
Hi all,
I am using Wavelet module for S-Plus to calculate
the wavelet functions. According to eq (2.2) on p14 in Bruce
and Gao's book (Applied Wavelet Analysis with S-Plus)
the following relationship should hold
phi_{jk} (t) = 2 ^(-j/2) phi(2^(-j)*t - k)
However, when I tried the following
> x0_plot(wavelet("d6",mother=F,shift=0,level=0),plot=F,J=2)
> x1_plot(wavelet("d6",mother=F,shift=0,level=-1),plot=F,J=2)
> x2_plot(wavelet("d6",mother=F,shift=0,level=-2),plot=F,J=2)
> x3_plot(wavelet("d6",mother=F,shift=0,level=-3),plot=F,J=2)
> x1$y/x0$y
[1] NA NA 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 NA
NA
> x2$y/x0$y
[1] NA NA 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 NA
NA
> x3$y/x0$y
[1] NA NA 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 NA
NA
In other words, the above experiments seem to indicate that
phi_{jk} (t) = 2 ^(-j) phi(2^(-j)*t - k)
(i.e. 2 ^(-j/2) is actually 2 ^(-j) in the above calculation)
Does anybody know what's going on here?
Cheers,
Shinn-Juh Lin
School of Finance & Economics
University of Technology, Sydney
[Answer WD 8.2 #15]
Date: Mon, 25 Jan 1999 09:17:16 +0100
From: Patrick Fischer <fischer@math.u-bordeaux.fr>
Subject: #30 Question: Wavelet Element Method
Dear readers,
I would need references about real numerical applications
of Wavelet Element methods for solving PDEs.
I'll summarize the answers and resubmit them to the Digest.
Regards,
Patrick FISCHER
Phone: (33 - 5) 57 96 21 30
Fax : (33 - 5) 56 84 26 26
http://www.math.u-bordeaux.fr/~fischer
Date: Mon, 25 Jan 1999 21:42:45 -0500
From: "Corey Albert Graves" <cagrave@eos.ncsu.edu>
Subject: #31 Question: Wavelets and classification
What feature, or set of features of signal segments will be good to
classify them into different classes that will be processed by
different wavelet bases (one wavelet for each class of signal
segments). The problem that I am addressing is this:
Given a finite set of N DWTs (different QMF filters), is there a way
to analyze some features of an abritrary signal segment and predict
which of the DWTs would be better at compressing the given signal
segment, without having to peform N different DWT-based compression
operations on the segment and comparing the results? If this is
feasible, then what features of the signal segment might be used to
make such a prediction? I anyone knows of any work that has been done
along these lines, please let me know.
Date: Tue, 26 Jan 1999 09:32:24 -0500 (EST)
From: Scott Spiegelberg <spiegel@theory.esm.rochester.edu>
Subject: #32 Question: Filters for music perception
Hello. I am a Ph.D candidate in music theory, specializing in the
psychoacoustics of music perception. I have been using Wavelet
Packets to analyze the attack transients of musical sounds. I chose
Bi-orthogonal wavelet filters almost randomly for these studies. Can
someone tell me how the filters vary, and which filter would be most
appropriate for parsing a complex acoustic wave by frequency over
time?
Scott Spiegelberg
Eastman School of Music
spiegel@theory.esm.rochester.edu
<webmaster@wavelet.org>