[Home] [Back Issues] [Search] [Calendar] [Vol 1] [Vol 2] [Vol 3] [Vol 4] [Vol 5] [Vol 6] [Vol 7]

The Wavelet Digest, Volume 7, Issue 5 (May 26, 1998)




Today's Editor: Wim Sweldens
                Lucent Technologies, Bell Laboratories
                wim@bell-labs.com



Today's Topics:

     1: Book:     The World According to Wavelets, Second Edition
     2: Thesis:   Optimal Perfect Reconstruction Filter Banks
     3: Preprint: Stochastic Atomic Decompositions in a Wavelet Dictionary
     4: Preprint: The Marr Wavelet to Examine Tree Growth Variation
     5: Preprint: Fast Adaptive Wavelet Packet Image Compression
     6: Preprint: Wavelet analysis edge detection.
     7: Preprint: MAPS Multiresolution Adaptive Parametrization of Surfaces
     8: Preprint: Unifying Algorithm for Daubechies Wavelets Collection
     9: Preprint: Two preprints from I. Selesnics on multiwavelets.
    10: Preprint: Preprints from Q. Sun on asymptotic smoothness
    11: Software: A wavelet toolbox for Matlab
    12: Meeting:  Curves and Surfaces, St-Malo, July 1999
    13: Meeting:  Royal Society London Feb 1999
    14: Meeting:  Workshop on Wavelets in Biology and Medicine
    15: Meeting:  IEEE Information Theory Workshop 
    16: Job:      Research opportunities at NASA Dryden
    17: Job:      Postdoctoral Position at Rice
    18: Job:      PhD Studentships at Napier University, Edinburgh, UK
    19: Job:      Postdoctoral position at U. of Illinois
    20: Contents: J. Approx. Th. - apr and may 98 TOC
    21: Web:      List of home pages of wavelet researchers
    22: Answer:   Wavelets & super-resolution of 2d images (WD 7.4 #25)
    23: Answer:   Wavelets & super-resolution of 2d images (WD 7.4 #25)
    24: Answer:   Parametrization of wavelets (WD 7.3 #19)
    25: Answer:   Error in Matlab manual (WD 7.4 #26)
    26: Answer:   Wavelet Transformation in medical diagnosis (WD 7.4 #17)
    27: Answer:   Wavelet on medical diagnosis (WD 7.4 #17)
    28: Answer:   Mathematics of Wavelets (WD 7.4 #16)
    29: Answer:   Ridge and skeleton extraction (WD 7.4 #23) 
    30: Question: Rediscovering Haar wavelets & wavelet packets
    31: Question: Wavelets and Vibration Monitory.
    32: Question: Wavelets & Power - Quality or Surge Protection?
    33: Question: Wavelets & Neural Network characterisation
    34: Question: Negative wavelet coefficients
    35: Question: Jackson-type inequality (simultaneous approximation)
    36: Question: Burt-Adelson wavelet
    37: Question: Wavelet for surface reconstruction?
    38: Question: Wavelet analysis of biological systems
    39: Question: About Shift-Invariant Wavelet
    40: Question: Maximum Absolute Deviation?
    41: Question: Wavelet spectral slopes
    42: Question: Introduction to wavelets for non-mathematicians/engineers
     
Current number of members: 11992

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:

 Jun 1-12:  Multiscale approaches for PDE, Marseilles, France
 Jun 8-12:  Wavelets, Neural Networks, Fuzzy Logic, Chaos, UCLA
 Jun 22:    IEEE Multimedia compression, Santa Barbara, CA
 Jun 29-Jul 10: EMS Summer School on Wavelets, Paris
 Jul 8-10:  Nonlinear Modeling, Leuven, Belgium
 Jul 30-Aug 7: Workshop on Self-Similar Systems, Dubna, Russia
 Sep 28-30: Applied Computational Electromagnetics, Penn State
 Oct 4-9:   School on Wavelets in the geosciences, Delft, Netherl.
 Oct 5-8:   10th GAMM-Workshop on Multigrid Methods, Bonn, Germany
*Feb 24-25: Royal Society London, UK
*Feb 24-26: IEEE Information Theory Workshop, Santa Fe, NM
*Jul 1-7:   Curves and Surfaces, St-Malo, France








Date: Tue, 26 May 1998 10:38:46 -0400
From: Amie Casper <marketing@akpeters.com>
Subject: #1 Book: The World According to Wavelets, Second Edition

The World According to Wavelets, Second Edition
Barbara Burke Hubbard

Now available from A K Peters, Ltd.; service@akpeters.com
330 pages; Hardcover; ISBN 1-56881-072-5
List price $40.00

This best-selling book introduces a broad audience including
scientists and engineers working in a variety of fields as well as
mathematicians from other subspecialties to one of the most active new
areas of applied mathematics and the story of its discovery and
development.  Organized in hypertext fashion, the book tells a
story of scientific discovery with separate brief entries for
technical terms and explicit appendices in a section called Beyond
Plain English.

>From the Table of Contents: Fourier Analysis: A Poem Transforms Our
World; Seeking New Tools; A New Language Acquires a Grammar;
Applications; Beyond Wavelets; The Fourier Transform; The Convergence
of Fourier Series and the Stability of the Solar System; Computing
Fourier Coefficients with Integrals; The Fast Fourier Transform; The
Continuous Wavelet Transform; Orthogonality and Scalar Products;
Multiresolution; The Fast Wavelet Transform; Wavelets in Two
Dimensions; Pyramid Algorithms of Burt and Adelson; Multiwavelets; The
Heisenberg Uncertainty Principle and Time-Frequency Decompositions;
Probability, Heisenberg, and Quantum Mechanics; Wavelets and Pure
Mathematics; Wavelets and Vision; Which Wavelet; Different Transforms:
A Summary; Wavelets, Music, and Speech; The Lifting Scheme; Best
Basis; Mathematical Symbols; Integrals; The Sampling Theorem: A Proof;
The Fourier Transform of a Periodic Function; An Example of an
Orthonormal Basis and a Proof of Fourier=92s Result; Further Reading;
Wavelet Software and Electronic Resources.

A K Peters, Ltd., 63 South Ave., Natick, MA 01760
Ph: 508/655-9933, Fax: 508/655-5847
<http://www.akpeters.com>



Date: Wed, 6 May 1998 16:55:09 +0200
From: ilangko@tele.ntnu.no (Ilangko Balasingham )
Subject: #2 Thesis: Optimal Perfect Reconstruction Filter Banks

Title: On Optimal Perfect Reconstruction Filter Banks 
       for Image Compression 

Author: Ilangko Balasingham <ilangko@tele.ntnu.no>
Address: Department of Telecommunications
        Norwegian University of Science and Technology
        N-7034 Trondheim, Norway
Advisor: Professor Tor A. Ramstad <tor@tele.ntnu.no>    

Compression of images have important applications in storage,
transmission, and processing. One of the modules in a compression
scheme is the filter bank.  This thesis focuses on selection,
optimization, and frequency partitioning of filter banks for both
lossless and lossy image compression. All filter banks constructed in
this thesis belong to the biorthogonal, perfect reconstruction class.

The wavelet transforms have received much attention lately.  The
connection between wavelet transform and filter bank is
investigated. It is known that the discrete wavelet transforms is a
sub-class of filter banks. In the thesis it is shown that by relaxing
the maximum regularity constraint, usually imposed on wavelet
transforms, and instead of introducing a smoothness constraint for
synthesis lowpass filter, higher coding gain in terms of PSNR and
visual quality is obtained.  This is demonstrated for a (9/7) system
constructed on such premises which outcompetes the popular 9/7 wavelet
transform.  Hence, the maximum regularity constraint is irrelevant.
The proposed subband coding gain optimized filter banks are among the
best in the literature.

Frequency domain tiling is scrutinized, and it is found that very good
coding results are obtained by using an 8 by 8 parallel filter bank
where the lowpass-lowpass band is further split by a dyadic filter
bank. The obtained coding results are among the best published so far
the literature.
 
A novel way of constructing integer coefficient filter banks is
proposed for lossless image coding. Discrete search of integer filter
coefficients under constrained optimization is used as a tool.  The
compression ratios for a number of images indicate that the tested
filter banks behave differently due to local variation of the image
statistics, and therefore a clear winner is difficult to choose.
However, a dyadic split using 2-channel 17/11 and 5/3 filter banks
have satisfactory performances. The compression ratios are comparable
to state-of-the-art lossless schemes when incorporating arithmetic
coder allocation as the compression engine.

A copy of the thesis can be obtained from 
http://www.tele.ntnu.no/~ilangko/thesis.html.



Date: Mon, 20 Apr 1998 11:29:43 +0100 (BST)
From: mafs <t.sapatinas@bristol.ac.uk>
Subject: #3 Preprint: Stochastic Atomic Decompositions in a Wavelet Dictionary

Authors: Felix Abramovich (University of Tel Aviv, Israel)
         Theofanis Sapatinas (University of Kent, United Kingdom)
         Bernard Silverman (University of Bristol, United Kingdom)

Title: Stochastic Atomic Decompositions in a Wavelet Dictionary

Abstract:

We consider random functions defined by atomic decompositions in a
wavelet dictionary.  The scale and dilation of the wavelet atoms are
not dyadic constraints, but the function is modelled as a sum of
wavelet functions at arbitrary positions and scales.  The locations of
the wavelet atoms and the magnitudes of their coefficients are chosen
with respect to a certain marked Poisson process model, allowing
intuitive notions about the functions genuinely to be modelled.  We
explore the properties of realizations from the proposed models by
investigating the relationships between the parameters of the models
and the Besov spaces within which realizations fall.  Such relations
may be seen as giving insight into the meaning of the Besov space
parameters themselves.  Moreover, although we do not elaborate on
this, the established relations make it possible in principle to
incorporate prior knowledge about the function's regularity into
statistical applications of the proposed models.

Keywords: Atomic Decompositions; Besov Spaces;
Poisson Processes; Wavelet Dictionaries.

The preprint is available at:

 http://www.stats.bris.ac.uk/~mafs/atomic4.ps.gz



Date: Mon, 20 Apr 1998 20:37:18 EDT
From: JohnsonP12 <johnsonp12@aol.com>
Subject: #4 Preprint: The Marr Wavelet to Examine Tree Growth Variation

Subject: Using the Marr Wavelet Function to Examine Tree Growth Variation

Author: Paul Johnson
             
Title: Software to Calculate the Power Spectrum Using the Marr Wavelet 
       Function to Examine Tree Growth Variation

e-mail: JohnsonP12@aol.com

Abstract:
              
Wavelets are mathematical functions that can be used to represent
environmental data. Wavelets are extremely useful when the data
contain sharp discontinuities. A repeated measures design with a
time-varying covariate is carried out to study the relationship
between environmental stress and tree growth. The distribution of
foliage, stem diameter, tree height and canopy biomass are measures
for tree growth. Environmental stress is measured through the use of
daily maximum and minimum air temperature, precipitation, percent of
cloud cover, maximum ozone concentration, soil moisture, soil
chemistry, percent shading and a crowding measure. The study
investigates the relationship between the stress events and tree
growth measurements.  A coefficient of dissimilarity that averages
assessments of diversity is used to describe size growth variation
among neighboring trees. The dissimilarity coefficient is computed
from pairs of trees and averages coefficients of variation. The data
contain sharp discontinuities, which leads to the use of wavelets to
examine tree growth variation. The Marr wavelet, also called the
Mexican hat wavelet, is obtained by calculating the second derivative
of a Gaussian. The author provides the wavelet algorithm to process
the data at different scales. The program calculates the wavelet
spectrum of the series, using the Marr as the mother wavelet. The
Fourier Transform (FT) of the Marr wavelet is normalized and then
multiplied by the FT of the time series.  The inverse FT is then used
to move the Wavelet Transform from Fourier space to real space. The
software can be changed to cover the option of using other wavelets
[e.g., Paul (m=4), Morlet or Derivative of a Gaussian DOG (m=6)].

Keywords: Wavelet transform, Fourier transform, Scale, Power spectrum, 
          Tree growth, Environmental Stress. 

The software is available from the author at   
http://members.aol.com/JohnsonP12/page2.html   



Date: Tue, 21 Apr 1998 13:33:14 -0400 (EDT)
From: "Francois G. Meyer" <meyer-francois@cs.yale.edu>
Subject: #5 Preprint: Fast Adaptive Wavelet Packet Image Compression

Title: Fast  Adaptive Wavelet Packet Image Compression

Author: Francois G. Meyer, Amir Z. Averbuch, Jan-Olov Stromberg

Available: noodle.med.yale.edu/~meyer/profile.html 
           - postscript file of the preprint
           - postscript file of the figures
           - software

Abstract: In this paper we address the problem of efficiently encoding
textured images. We propose a fast wavelet packet compression
algorithm.  The wavelet packet expansion exploits factorized
non-separable very fast filters, and is faster by a factor of 4 than a
regular convolution and decimation of the image with separable
one-dimensional filters. The selection of the best basis does not
require any iterative procedure. This algorithm provides a very
efficient implementation of the best basis algorithm. The wavelet
packet coefficients are organized by increasing frequency into
sequences of rapidly decreasing coefficients. The coefficients are
quantized with uniform quantization, and bitplane encoded.  The
evaluation of the algorithm was performed on a large class of textured
images. The results indicate that this new fast wavelet packet coder
is competitive to, and often outperforms the best wavelet coding
algorithms, both visually and in term of the quadratic error.

An implementation of the algorithm is available on the web site.

    Francois G. Meyer                             francois.meyer@yale.edu

Yale University School of Medicine,                   tel:(203) 737 6037
Department of Diagnostic Radiology                        (203) 785 2427 
333 Cedar Street, P.O. Box 208042                     fax:(203) 737 4273     
New Haven, CT 06520-8042, USA                                           

             http://noodle.med.yale.edu/~meyer/profile.html



Date: Wed, 22 Apr 1998 17:33:54 +0900
From: applma@cqu.edu.cn
Subject: #6 Preprint: Wavelet analysis edge detection.

          Detect edges respectively from the image and projections
              of ICT with wavelet analysis method

               Li Zeng     Xu Wen-Zhi      Chen Ting-Huai
    Dept. of applied math., Chongqing University, Chongqing, 400044,
    P. R. China  E-mail   applma@cqu.edu.cn

Abstract In industrial computerized tomography ( I C T ) , with the
wavelet analysis method, we detect edges within a cross section from
the image or projections. This method can detect edges from the noised
image of ICT.  Comparing with the Laplace method that detect edges
from the noised image of ICT, this method can reduce the noise of the
edge image and can filter the non-white noise. These methods can be
applied to medical computerized tomography, too.



Date: Wed, 29 Apr 98 14:58:02 EDT
From: Wim Sweldens <wim>
Subject: #7 Preprint: MAPS Multiresolution Adaptive Parametrization of Surfaces

Title: MAPS: Multiresolution Adaptive Parameterization of Surfaces

Authors: Aaron W. F. Lee, Wim Sweldens, Peter Schroeder, Lawrence
         Cowsar, David Dobkin  

Abstract: We construct smooth parameterizations of irregular
connectivity triangulations of arbitrary genus 2-manifolds. Our
algorithm uses hierarchical simplification to efficiently induce a
parameterization of the original mesh over a base domain consisting of
a small number of triangles. This initial parameterization is further
improved through a hierarchical smoothing procedure based on Loop
subdivision applied in the parameter domain. Our method supports both
fully automatic and user constrained operations. In the latter, we
accommodate point and edge constraints to force the alignment of
iso-parameter lines with desired features. We show how to use the
parameterization for fast, hierarchical subdivision connectivity
remeshing with guaranteed error bounds. The remeshing algorithm
constructs an adaptively subdivided mesh directly without first
resorting to uniform subdivision followed by subsequent
sparsification. It thus avoids the exponential cost of the latter. Our
parameterizations are also useful for texture mapping and morphing
applications, among others.

Status: Computer Graphics Proceedings (SIGGRAPH 98), to appear, 1998. 

Download Postscript or PDF at

/who/wim/papers/sig98/
http://www.multires.caltech.edu/pubs/
or
http://www.cs.princeton.edu/~wailee/maps.html



Date: Fri, 08 May 1998 12:31:51 -0700
From: Carl Taswell <taswell@toolsmiths.com>
Subject: #8 Preprint: Unifying Algorithm for Daubechies Wavelets Collection

Title: A Spectral-Factorization Combinatorial-Search Algorithm
Unifying the Systematized Collection of Daubechies Wavelets

Author: Carl Taswell

In press: Invited Paper and Session Keynote Lecture, Intl Conf on
Systems, Signals, Control, Computers to be held Durban South Africa,
September 1998.

Available: www.toolsmiths.com/papers.shtml

Abstract: A spectral-factorization combinatorial-search algorithm has
been developed for unifying a systematized collection of Daubechies
minimum length maximum flatness wavelet filters optimized for a
diverse assortment of criteria. This systematized collection comprises
real and complex orthogonal and biorthogonal wavelets in families
within which the filters are indexed by the number of roots at
$z=-1$. The main algorithm incorporates spectral factorization of the
Daubechies polynomial with a combinatorial search of spectral factor
root sets indexed by binary codes.  The selected spectral factors are
found by optimizing the desired criterion characterizing either the
filter roots or coefficients.  Daubechies wavelet filter families have
been systematized to include those optimized for time domain
regularity, frequency domain selectivity, time frequency uncertainty,
and phase nonlinearity. The latter criterion permits construction of
the orthogonal least and most asymmetric real and least and most
symmetric complex filters. Biorthogonal symmetric spline and balanced
length filters with linear phase are also computable by these methods.



Date: Fri, 8 May 1998 20:45:24 -0400 (EDT)
From: Ivan Selesnick <selesi@taco.poly.edu>
Subject: #9 Preprint: Two preprints from I. Selesnics on multiwavelets.

1:

Cardinal Multiwavelets and the Sampling Theorem

This paper considers the classical Shannon sampling theorem in
multiresolution spaces with scaling functions as interpolants.  As
discussed by Xia and Zhang, for an orthogonal scaling function to
support such a sampling theorem, the scaling function must be
cardinal.  They also showed that the only orthogonal scaling function
that is both cardinal and of compact support is the Haar function,
which has approximation order 1 only.  This paper addresses the same
question, but in the multiwavelet context, where the situation is
different.  This paper presents the construction of orthogonal
multiscaling functions that are simultaneously cardinal, of compact
support, and have approximation order K>1.  The scaling functions
thereby support a Shannon-like sampling theorem.

2:

Multiwavelets with Extra Approximation Properties

Because multiwavelet bases normally lack important properties
traditional wavelet bases (of equal approximation order)
possess, the discrete multiwavelet transform is less useful
for signal processing, unless preceded by a preprocessing step.
This paper examines properties of orthogonal multiwavelet bases,
with approximation order K>1, that possess those properties
normally absent.  For these ``balanced'' bases (after Lebrun and
Vetterli) prefiltering can be avoided.  Using results regarding
M-band wavelet bases, it has been found that balanced multiwavelet
bases can be characterized in terms of the divisibility of certain
transfer functions by powers of (z^{-2r}-1)/(z^{-1}-1).  The paper
also presents a balanced version of the DGHM basis  --- the scaling
functions are simultaneously symmetric, orthogonal and of compact
support.

Ivan Selesnick
Polytechnic University
Brooklyn, NY
selesi@taco.poly.edu

http://taco.poly.edu/selesi/



Date: Wed, 20 May 1998 20:52:49 +0800
From: "Dr. Sun Qi yu" <matsunqy@leonis.nus.edu.sg>
Subject: #10 Preprint: Preprints from Q. Sun on asymptotic smoothness

Dear editors of Wavelet Digest,

I enclosed three preprints on asympotic estimate of Sobolev exponent of
some refinable functions.

1. Title: Asymptotic Regularity of Daubechies' Scaling  Functions
   Authors: K.-S. Lau and Qiyu Sun

2. Title: Regularity of Butterworth Refinable Functions
   Authors: Ai Hua Fan and Qiyu Sun

3. Title: Sobolev Exponent Estimate and Asymptotic Regularity of M
   Band Daubechies' Scaling Functions
   Author: Qiyu Sun

In our papers, we introduce a new method to direct estimate Sobolev
exponent and apply it to the asymptotic estimate of Sobolev exponent
of M band Daubechies' scaling functions and Butterworth refinable
functions.

For the Daubechies' scaling functions, the previous difference between
upper and lower estimate of Sobolev exponent $s_p, 1\le p\le \infty$
obtained by Volker (IEEE Trans. Inform. Th., 38(1992), 872-876),
independently by Cohen and Conze (Revista Mat. Iberoamericana,
8(1992), 527-591) is about $C ln N$ for $M=2$, where $N$ is the
papamater in the construction of Daubechies' scaling functions. The
difference obtained by Bi, Dai and Sun (Construction of compactly
supported M band wavelets, Applied Computational Harmonic Analysis, to
appear) is a bounded constant for $M\ge 2$.

In our papers, we introduce a term independent of $0<p\le \infty$
which has explicit expression and show that the difference between the
term above and the Sobolev exponent $s_p, 0<p\le\infty$ of Daubechies'
scaling functions is about $C r^N$ for some constants $C$ and $0<r<1$
independent of the parameter $N$. This also affirms the phenomenon
observed by Cohen and Daubechies (Revista Mat. Iberoamericana,
12(1996), 527-591). The phenomenon is observed by Cohen and Daubechies
through numerical computation for $M=2$. In particular, it is their
paper that inspires us to consider the asymptotic estimate of Sobolev
exponent.

About Butterworth refinable functions, Cohen and Daubechies shows in
their paper (Revista Mat. Iberoamericana, 12(1996), 527-591) by
numerical computation that its Sobolev exponent has the similar
phenomenon as the one of Daubechies' scaling functions. Because the
Butterworth refinable functions have infinite support, the phenomenon
is not so clear as the one of Daubechies' scaling functions.  In our
paper, we give the upper bound $N\ln3/\ln2+\ln(1+3^{-N})/\ln 2$ and $
N\ln3/\ln2-\ln(1+r^N)/\ln 2$ of Sobolev exponent $s_p, 0<p\le \infty$
of Butterworth refinable functions and affirm the phenomenon for
Butterworth refinable functions, where $0<r<1$ and $N$ is the
parameter in the construction of Butterworth refinable functions. By
our computation, $r$ can be chosen as $0.9787028^p$.

The second and third paper can be download from
 
http://haar.math.nus.edu.sg/~xiatao/sun/SQY1097.ps.Z

http://haar.math.nus.edu.sg/~xiatao/sun/Mregular.ps.Z

With best wishes

Qiyu Sun

Email Address: matsunqy@leonis.nus.edu.sg
Fax: 65-7995452
Homepage: http://haar.math.nus.sg/qsun



Date: Fri, 1 May 1998 00:01:22 -0400 (EDT)
From: Harri Ojanen <ojanen@math.rutgers.edu>
Subject: #11 Software: A wavelet toolbox for Matlab

I would like to announce the availability of a wavelet toolbox for
Matlab that implements the following wavelet and wavelet packet
algorithms:

    One- and two-dimensional (periodic) fast wavelet and wavelet
    packet transforms and the best basis algorithm for wavelet packets.

    An implementation of the fast matrix multiplication algorithm of
    Beylkin, Coifman, and Rokhlin (Comm. Pure Applied Math 44(2), 1991)
    for both wavelets and wavelet packets.

    Various demonstrations on visualizing wavelets, signal 
    analysis, and the multiplication algorithm.

The programs in this toolbox are mostly intended for learning about
wavelets. The toolbox grew out of my desire to teach myself about
wavelets and now I feel it has matured to the point where it might
prove useful to others. I think some of the visualization tools and
demos are rather unique.

All comments, criticism, bug reports, etc. are certainly welcomed!

The programs and documentation (both in Postscript and online)
are available on the web at

        http://www.math.rutgers.edu/~ojanen/wavekit/

System requirements: Matlab 5.1 or later and a compatible C-compiler.

Harri Ojanen
ojanen@math.rutgers.edu
http://www.math.rutgers.edu/~ojanen/



Date: Fri, 24 Apr 1998 17:20:36 METDST
From: Christophe Rabut <rabut@gmm.insa-tlse.fr>
Subject: #12 Meeting: Curves and Surfaces, St-Malo, July 1999

The organizers are pleased to inform you the following conference :
  
  (more details will be sent later on ; please look at our Web page)
  
  Fourth AFA International Conference on
    
         "CURVES AND SURFACES"
    
    Palais du Grand Large,
    Saint-Malo (North-West of France), July 1-7, 1999.
    
    http://www.enst.fr/~afa/saint-malo/
    Email: saint-malo@imag.fr
    Fax: 33 76 63 12 63
    Postal address: "Curves and Surfaces", LMC-IMAG,
    BP 53, 38041 Grenoble, cedex 09, France).
    
  ORGANIZERS:
    Albert COHEN (Paris, France),
    Pierre-Jean LAURENT (Grenoble, France),
    Christophe RABUT (Toulouse, France),
    Paul SABLONNIERE (Rennes, France),
    Larry SCHUMAKER (Nashville, USA).
    
  SCIENTIFIC COMMITTEE:
    Martin Buhmann (Germany), Rosemary Chang (USA), Wolfgang Dahmen (Germany),
    Carl de Boor (USA), Ron DeVore (USA), Nira Dyn (Israel), Rida Farouki
    (USA), Mariano Gasca (Spain), Ron Goldman (USA),Tim Goodman (UK), Josef
    Hoschek (Germany), Kurt Jetter (Germany), Yvon Maday (France), Alain Le
    Mehaute (France), Will Light (UK), Tom Lyche (Norway), Yves Meyer (France),
    Gregory Nielson (USA), Allan Pinkus (Israel), Helmut Pottmann (Austria),
    Malcolm Sabin (UK), Robert Schaback (Germany), Hans-Peter Seidel (Germany).
    
  ORGANIZING COMMITTEE:
    Luca Amodei, Jean-Luc Bauchat, Jean-Charles Fiorot, Jean Gaches,
    Georges-Pierre Bonneau, Yvon Lafranche, Marie-Laurence Mazure, Jean-Louis
    Merrien, Christine Potier, Christine Vercken.
    
  CONFERENCE TOPICS:
    - Representation and approximation of curves and surfaces (splines, finite
    elements, subdivision,...),
    - Computer-aided geometric design,
    - Interpolation and smoothing (scattered data, variational methods,...),
    - Multiresolution techniques (wavelets, hierarchical basis,...),
    - Curves and surfaces in industry,
    - Related applications (image analysis and synthesis, robotics,
    visualization, reverse engineering, CAD-CAM,...).
    
  INVITED SPEAKERS:
    Zbigniew CIESIELSKI (Poland), Ingrid DAUBECHIES (USA), David DONOHO (USA),
    Rida FAROUKI (USA), Yvon MADAY (France), Helmut POTTMANN (Austria), Alyn
    ROCKWOOD (USA), Amos RON (USA), Robert SCHABACK (Germany), Wim SWELDENS
    (USA).
    
  MINI-SYMPOSIUM  ORGANIZERS
    Wolfgang DAHMEN (Germany), Ron DeVORE (USA), David DONOHO (USA), Josef
    HOSCHEK (Germany), Bernard LACOLLE (France), Helmut POTTMANN (Austria),
    Malcolm SABIN (UK), Joachim STOECKLER (Germany).
    


Date: Fri, 1 May 1998 21:42:23 +0100 (BST)
From: Bernard Silverman <b.w.silverman@bristol.ac.uk>
Subject: #13 Meeting: Royal Society London Feb 1999

Wavelets: the key to intermittent information

Royal Society, London, 24 and 25 February 1999

This Royal Society Discussion Meeting will survey the application and
potential of wavelet methods in a very diverse range of fields,
drawing connections with physical systems, turbulence, statistics,
mechanical engineering, neural networks, physiology, vision
engineering, economics, signal processing, astronomy and other areas.
It is expected that the speakers will be Alain Arneodo, Emmanuel
Candes & David Donoho, Marie Farge, David Field, Alex Grossman, Iain
Johnstone, Nick Kingsbury, David Newland, Ue-Li Pen, James Ramsey,
Bernard Silverman, Wim Sweldens & Ingrid Daubechies, Christos
Vassilicos, and Martin Vetterli.

In addition there will be contributed poster sessions organised by
Irene Moroz and Guy Nason, and a number of public panel discussions
involving the main speakers and other active researchers.  A call for
submission of posters will be made later in 1998.

The meeting is intended both to be a showcase to a general scientific
audience of the scope of wavelet methods, and to provide stimulus and
direction to future research.  It should be of interest to
non-specialists and specialists alike.

The meeting will take place at the Royal Society in London.  There is
no registration fee, but you are asked to register in advance of the
meeting.  Registration details will be available later in 1998 from
http://www.royalsoc.ac.uk/st_mtgs.htm

Please note the dates for your diaries!

Bernard Silverman for the organizing committee (himself, Ingrid
Daubechies, Julian Hunt and Christos Vassilicos) 1 May 1998



Date: Fri, 15 May 1998 16:39:36 -0400
From: Akram Aldroubi <aldroubi@math.vanderbilt.edu>
Subject: #14 Meeting: Workshop on Wavelets in Biology and Medicine

Workshop on Wavelets in Biology and Medicine

Organizer: Professor Murray Eden MIT/NIH

The workshop on "Wavelets in Biology and Medicine" will precede The
20th Annual International Conference of the IEEE Engineering in
Medicine and Biology Society (EMBS'98)
(http://www.ee.cuhk.edu.hk/embs98.html).  It will be held during the
preconference period, 27-28 October, 1998, The Hong Kong Convention
and Exhibition Center, HONG KONG.

Speakers:
Akram Aldroubi, Vanderbilt University
Dennis Healy, Dartmouth College/ DARPA
Murray Eden, MIT/NIH

This workshop is a one day course.  It will provide a step-by-step
introduction to the wavelet transform emphasizing the properties that
are the most relevant for biomedical signal and image processing. It
will also review the most important applications of wavelet techniques
in medicine and biology. The workshops on wavelets will be tutorial
flavor, and will be addressed to participants who are not necessarily
wavelet experts. It will also provide a timely update on the most
recent developments in this area.

For more information you can contact Dr. Eden at meden@world.std.com

Akram Aldroubi, Associate Professor
Dept.of Math.
Vanderbilt Univ.
Nashville TN, 37209
Tel:US+(615)-322-6656
FAX:US+(615) 343-0215
Web Page: http://atlas.math.vanderbilt.edu/~aldroubi



Date: Wed, 20 May 1998 11:54:34 -0500 (CDT)
From: Pierre Moulin <moulin@ifp.uiuc.edu>
Subject: #15 Meeting: IEEE Information Theory Workshop 

        1999 IEEE Information Theory Workshop on Detection, Estimation,
                        Classification and Imaging (DECI)

                February 24---26, 1999, Santa Fe, New Mexico (USA)

                        FIRST CALL FOR PAPERS

Detection and estimation theory have historically been closely linked
to Information Theory.  Analysis of communication systems relies
heavily on and contributes to advances in detection and estimation
theory.  Considerable theoretical and practical advances in this area
have been made possible by the fostering of ideas from Statistics and
Information Theory.  This workshop will complement those activities by
seeking contributions from researchers in signal processing, image
processing, image understanding, pattern recognition, and
communication theory, whose work is heavily influenced by information
theoretic considerations.  While novel applications will play
important roles, new theoretical results are expected to dominate.

The workshop will feature three plenary speakers: Andrew Barron
(Yale), H. Vincent Poor (Princeton), and Michael I. Miller (Washington
University).  Invited talks and contributed talks will be presented in
the following areas:

                        Detection Theory
                        Estimation Theory
                        Classification
                        Statistical Imaging
                        Regularization for inverse problems
                        Random Processes
                        Inference based on compressed data
                        Signal Processing Applications of IT

Of particular interest are papers dealing with nonparametric and
robust methods, methods for non-euclidean spaces, alternating
maximization methods, high-dimensional inverse problems, and
dimensionality reduction.  Authors interested in submitting a
contribution should mail four copies of a 2-3 page summary to
Prof. J. O'Sullivan (address below) by October 9, 1998.  Papers will
be presented either as 20-minute talks or as posters.  One-page
abstracts of all papers will appear in the proceedings of the workshop
and will be posted to the workshop web site prior to the workshop.

The workshop will be held in historic Hotel Loretto, in the heart of
Santa Fe.  Santa Fe possesses a rich Spanish and Native American
culture and is located in the vicinity of excellent ski resorts.  It
is expected that a small number of student travel grants to the
workshop will be available. Detailed information will be included in
the final call for papers and will be posted to the workshop web site:

                        http://www.ifp.uiuc.edu/itw-deci

Inquiries about the workshop should be directed to one of the three co-chairs:

Prof. Alfred O. Hero III  Prof. Pierre Moulin     Prof. J. A. O'Sullivan
EECS Department           ECE Department          EE Department
U. of Michigan            U. of Illinois          Washington University
1301 Beal Avenue          Beckman Institute       Campus Box 1127
Ann Arbor, MI 48109       405 N. Mathews Ave.     St. Louis, MO  63130
hero@eecs.umich.edu       Urbana, IL 61801        jao@ee.wustl.edu
                          moulin@ifp.uiuc.edu



Date: Tue, 28 Apr 1998 11:04:13 -0700 (PDT)
From: Marty Brenner <gonzo@xrd.dfrc.nasa.gov>
Subject: #16 Job: Research opportunities at NASA Dryden

Contributed by: Martin J. Brenner

                NASA/RRA/NRC Research Awards for 1998
                at NASA Dryden Flight Research Center

NASA Resident Research Associateship (RRA)
administered by the National Research Council (NRC)
Postdoctoral and Senior Research Awards for 1998
Opportunity for Research at NASA Dryden Flight Research Center

Proposed Research Title: 
Nonlinear Estimation of Aerospace Systems using Multiscale System Theory

The ability to identify and control nonlinear phenomena is extremely
important to the aerospace community. Understanding aeroelastic
nonlinear dynamics is becoming vital for advanced design
configurations.  Stability estimation from flight data must account
for nonlinearity to distinguish linear from nonlinear dynamics for
model validation.

Confidence bounds are desired in identification schemes to guarantee
system stability within a flight regime. Identification from flight
data consistent with the uncertainty and disturbance models is desired
for an explicit measure of stability at any flight condition.

Application of time-frequency multiscale methods is an attractive
approach to this research. Wavelets provide systematic tools for
signal processing of nonlinear and time-varying systems. Wavelet
research at NASA Dryden concentrates on algorithms to generate
time-frequency multiscale representations to identify nonlinear and
time-varying aspects of the system dynamics.  Multiscale system theory
may be a promising approach to this research.

Dryden is studying linear and nonlinear identification algorithms and
methodologies for in-flight aeroservoelastic robust stability
determination.  Intelligent data processing and analysis methods are
desired for accurate control-oriented stability determination during
flight testing.  This integrated research effort requires innovative
data processing, multiscale identification, and robust system theory
applications.

Contact:                Marty Brenner
                        NASA Dryden Flight Research Center
                        MS 4840D / RS
                        Edwards, California 93523-0273

                        Phone: 805-258-3793
                        Fax: 805-258-2586
                        email: martin.brenner@dfrc.nasa.gov

Postdoctoral Research Associateships are awarded to persons who have
held the doctorate less than five years and are made initially for one
year. Extensions to two years are common.

Senior Research Associateships are awarded only to applicants who have
held the doctorate five years or more and demonstrate research
experience that has resulted in significant contributions and
recognition as established investigators. They are usually for one
year, but awards for periods of three months or longer may be
considered. A Research Associate must conduct his research full-time
and on-site (NASA Dryden).

This opportunity for research IS ONLY OPEN TO US CITIZENS.

The FINAL review of applications is conducted by special panels
appointed by the NRC. Evaluations are conducted in February, June, and
October.  Awards are made only by the NRC. The current annual stipend
is approximately $50,000 for Postdoctoral Research Associate, and
approximately $80,000 for Senior Research Associate.

Detailed information on procedures and all necessary application
materials and supporting documents are available on request from:

                Associateship Programs          E-mail: rap@nas.edu
                TJ 2114                         WWW: http://www.nas.edu/rap
                National Research Council       Gopher: nas.edu/rap
                2101 Constitution Avenue NW
                Washington, D.C. 20418



Date: Mon, 4 May 1998 13:39:11 -0500
From: "Jun Tian" <juntian@rice.edu>
Subject: #17 Job: Postdoctoral Position at Rice

Postdoctoral Research Position
Computational Mathematics Laboratory, Rice University

Applications are solicited for a postdoctoral position in the
Computational Mathematics Laboratory at Rice University.  The
applicant should have a doctorate in mathematics, applied mathematics,
or electrical engineering (signal processing) with a strong
mathematics background.  In particular, a background in wavelet
analysis and/or wavelet signal processing is essential. Matlab is an
essential tool that will be necessary, and additional programming
languages (such as C, C++, etc.) would be useful.  The problems
involve applications of wavelets to pattern recognition and image
processing issues involving sonar and synthetic aperture radar imagery
data, as well as medical and geophysical imaging problems. The
Computational Mathematics Laboratory is a part of the Computer and
Information Technology Institute at Rice and involves faculty,
researchers and students in the departments of mathematics and
electrical and computer engineering.  Applicants should send their
resumes with a list of three references and any other supporting
documents (e.g. publications) before 1 June 1998 to:

Professor R. O. Wells, Jr.,
Director, Computational Mathematics Laboratory
Department of Mathematics
Rice University
6100 Main Street
Houston, TX 77005

Electronic submissions should be sent to: cml@math.rice.edu

Fax applications could be sent to: 713-285-5437



Date: Thu, 7 May 1998 14:27:26 +0100
From: p.addison@napier.ac.uk
Subject: #18 Job: PhD Studentships at Napier University, Edinburgh, UK

     A number of PhD studentships will be available within the
     Department of Civil and Transportation Engineering, at Napier
     University, Edinburgh, starting in September of this year. Two
     topics are on offer, one specifically looking at fractal and
     wavelet analysis of crack patterns the other of a more general
     nature, these are:
     
     1 - CHARACTERISATION OF CONCRETE CRACKING USING BOTH WAVELET
     TRANSFORM AND FRACTAL GEOMETRIC ANALYSES: This project aims to
     extend current work in this area and use both fractal geometric
     techniques and wavelet analysis to characterise both crack
     profiles and internal crack surfaces. The work seeks to develop
     new methods of non destructive testing using crack
     characteristics. The fractal work involves the geometric
     description of the crack geometry in terms of an effective
     Fokker-Planck equation. The wavelet work involves the use of a
     variety of continuous and discrete transforms to both grade the
     material and locate areas where subsurface inhomogenieties may
     exist.
     
     2 - WAVELET TRANSFORM ANALYSIS IN CIVIL ENGINEERING: This project
     aims to investigate the use of the wavelet transform, in a
     variety of civil engineering signal analysis applications
     including non-destructive testing signals; rainfall times series;
     turbulence data; crack pattermns, and ocean surface drifter
     trajectories. The focus of the work will very much depend upon
     the interest(s) of the applicant.
     
     For background information on related research topics see website:
     
     http://www.napier.ac.uk/depts/cte/staff/paddison.html
     
     For more specific details contact
     
     Dr. P.S. Addison
     Department of Civil and Transportation Engineering 
     Napier University
     Merchiston Campus
     10 Colinton Road
     Edinburgh
     EH10 5DT
     Scotland
     
     Tel    +44 (0)131 455 2552
     Fax    +44 (0)131 455 2239
     e-mail p.addison@napier.ac.uk



Date: Sat, 9 May 1998 16:07:50 -0500 (CDT)
From: Pierre Moulin <moulin@ifp.uiuc.edu>
Subject: #19 Job: Postdoctoral position at U. of Illinois

                     POSTDOCTORAL POSITION
IN STATISTICAL SIGNAL AND IMAGE PROCESSING AND IMAGING SCIENCE

A postdoctoral position in statistical signal and image processing and
imaging science is available at the Coordinated Science Laboratory of
the University of Illinois at Urbana-Champaign.  The project aims at
developing novel concepts in imaging with emphasis on modeling of
image sources and imaging systems.  It will involve a synthesis of
nonlinear inverse problem theory with statistical inference
techniques.  A strong background in detection/estimation and random
processes is a must.  Desirable background includes statistical
pattern recognition, information theory, optimization, imaging
systems, image processing, inverse problem theory, wavelets, and
computational algorithms.  A background in electromagnetic waves,
scattering, and propagation is an advantage.

Prospective candidates should send CV and names and emails of three
references to Ms. Nancy Carr, n-carr@uiuc.edu

Yoram Bresler                               Pierre Moulin
http://what.csl.uiuc.edu/~yoram             http://www.ifp.uiuc.edu/~moulin

                     Coordinated Science Laboratory
                 University  of Illinois at Urbana-Champaign



Date: Thu, 23 Apr 1998 20:21:02 -0400 (EDT)
From: Thomas Hogan <hogan@math.ohio-state.edu>
Subject: #20 Contents: J. Approx. Th. - apr and may 98 TOC

Table of Contents: J. Approx. Theory, Volume 93, Number 1, April 1998

R. J. Kooman
Asymptotic behaviour of solutions of linear recurrences and sequences of
   M\"obius-transformations
1--58

Suresh Kumar Lodha and Ron Goldman
Lattices and algorithms for bivariate Bernstein, Lagrange, Newton, and
   other related polynomial bases based on duality between $L$-bases and
   $B$-bases
59--99

Xin Min Li
Uniform bounds for sampling expansions
100--113

E. S. Belinsky
Estimates of entropy numbers and Gaussian measures for classes of
   functions with bounded mixed derivative
114--127

Jos\'e A. Adell and Ana P\'erez-Palomares
Best constants in preservation inequalities concerning the first modulus
   and Lipschitz classes for Bernstein-type operators
128--139

Ingrid Carbone
Shape preserving properties of some positive linear operators on
   unbounded intervals
140--156

Geno Nikolov
On certain Duffin and Schaeffer type inequalities
157--176

Table of Contents: J. Approx. Theory, Volume 93, Number 2, May 1998

D. Barrios Rolan\'{\i}a, G. L\'opez Lagomasino,
   A. Mart\'{\i}nez Finkelshtein, and E. Torrano Gim\'enez
On the domain of convergence and poles of complex $J$-fractions
177--200

Ping Zhou
Multivariate Pad\'e approximants associated with functional relations
201--230

Du Jinyuan
Quadrature formulas of quasi-interpolation type for singular integrals
   with Hilbert kernel
231--257

Holger Wendland
Error estimates for interpolation by compactly supported radial basis
   functions of minimal degree
258--272

R. A. DeVore, G. C. Kyriazis, and P. Wang
Multiscale characterizations of Besov spaces on bounded domains
273--292

Lawrence A. Harris
Coefficients of polynomials of restricted growth on the real line
293--312

Mohammad Kazim Khan and Muhammad Hanif
On some refinements of Jensen's inequality
313--330

Jos\'e Mendoza
Proximinality in $L_p(\mu,X)$
331--343

Note

Guo Tie-Xin and You Zhao-Yong
A note on pointwise best approximation
344--347



Date: Fri, 8 May 1998 15:48:59 +0200 (MET DST)
From: Andreas Klappenecker <klappi@ira.uka.de>
Subject: #21 Web: List of home pages of wavelet researchers

Hello,

some time ago I started to assemble a small list of people working in
wavelet analysis:
      http://iaks-www.ira.uka.de/home/klappi/people.html
You can help me to improve this list by sending me missing home pages
(just drop me a short e-mail containing the name of the researcher
and URL of the home page).

Thank you!
      Andreas Klappenecker
      (e-mail: klappi@ira.uka.de)



Date: Wed, 22 Apr 1998 14:01:58 -0400 (EDT)
From: William Knox Carey <knox@anise.ee.cornell.edu>
Subject: #22 Answer: Wavelets & super-resolution of 2d images (WD 7.4 #25)

Hi Nick:

  I've done some work with Professor Sheila Hemami at Cornell on
"super-resolution" (i.e. interpolation) of images with wavelets. You can
find a preprint of our paper on that topic at:
http://tweed.ee.cornell.edu/~knox/research.html
The paper is called "Regularity-Preserving Image Interpolation".

-knox



Date: Sat, 18 Apr 1998 14:26:21 +0800
From: Bernard Cena <bernard@cs.uwa.edu.au>
Subject: #23 Answer: Wavelets & super-resolution of 2d images (WD 7.4 #25)

Nick Pelling writes:

Has anyone seen any references to wavelets' application to 2d-image
super-resolution? ie, using a wavelet decomposition of an image to
predict higher-order wavelets, and thus infer likely (but
non-existent) detail. 8^)

Answer:

I don't recall any direct references to super-resolution, but as far
as I understand you want to interpolate in-between pixels of your
image ?  As you are no doubt aware, when reconstructing data with the
Discrete Wavelet Transform (DWT) if you cut out all wavelet
coefficients leaving only scaling functon coefficients you will
basically SEE the shape of the scaling functions in your
reconstruction. This is the reason why people use smooth fiters for
reconstruction of "denoised" or "compressed" images because cutting
out "details" results in showing the characteristics of the scaling
function and wavelet- if that scaling function/wavelet is very
"jagged" ie. not smooth, you will get ugly artifacts in your
reconstruction. The same applies if you try to do super-resolution in
images which is basically the same as interpolating a larger image
from a smaller one.

I am not sure what you mean by "higher order" wavelets - I presume you
meant finer scale scaling functions ? Anyway, I myself am working on
the application of wavelets to data visualisation using volume
rendering.  To do that I use Wim Sweldens' Lifting Scheme to arrive at
arbitary order interpolating and average interpolating scaling
functions/wavelets.  The trick in all this is to realise what
"continuous function" does your date "come from", or more precisely,
has been sampled from. If you want to achieve "super-resolution" you
just keep on iterating the reconstruction phase of the DWT which will
give you proper results if you match the scaling function to the type
of image properly. In essence if comes down to identifying a
subdivision scheme which given samples of a function will result in
the same function when the subdivision algorithm is iterated.  This is
the beauty of wavelets: you can design your own wavelets for a
particular purpose.

So for example, CCD arrays (digital cameras) perform some sort of
averaging (integration) for each pixel. It would therefore make sense
to use some order of an average-interpolating subdivision scheme to
build your scaling function and associated wavelet. The lifting scheme
is ideal for this as it starts with a subdivision scheme and builds a
biorthogonal filter set with the advantages of being able to handle
boundaries, non- uniform sampling, non-Euclidean spaces (eg. wavelets
on the sphere), integer transform, in-place computation, etc.

Please refer to Wim Sweldens' bibliography at:
/who/wim/
and especially the paper "Building your own wavelets at home".
The interpolating subdivision scheme has been implemented
and can be obtained as LIFTPACK at:
http://www.cs.sc.edu/~fernande/liftpack/index.html

Good luck!
Bernard Cena

Bernard Cena (PhD student - Med Vis)  _--_|\    tel: +61 8 9380 3778
Department of Computer Science       /      \   fax: +61 8 9380 1089
The University of Western Australia  *_.--._/  home: +61 8 9337 8389
Nedlands, 6907, WA, Australia              v    mob: +61 417 923 534
http://www.cs.uwa.edu.au/~bernard ----- email: bernard@cs.uwa.edu.au



Date: Mon, 20 Apr 1998 21:36:59 +0200 (MET DST)
From: Algirdas Bastys <abba@ieva.maf.vu.lt>
Subject: #24 Answer: Parametrization of wavelets (WD 7.3 #19)

Answer to Question 19 "Parametrization of wavelets", WD, Vol.7, Nr 3

Dear Senthil,

To parametrize 2N-length filter h_n, n=0,...,2N-1, that correspond to
a tight (as a rule to orthogonal) scaling function $\phi$ you can
choose N-1 parameters 0\le\alpha_1,\alpha_2,\ldots,\alpha_{N-1}<1, and
use the following recursion:

1. Initialization

  h_0^0=1/\sqrt2; h_1^0=1/\sqrt2;
  g_0^0=-h_1^0; g_1^0=h_0^0;

2. Recursion

  h_i^k = cos\alpha_k (cos\alpha_k h_i^{k-1} + sin\alpha_k g_i^{k-1}) +
    sin\alpha_k (sin\alpha_k h_{i-2}^{k-1} - cos\alpha_k g_{i-2}^{k-1});
  g_i^k = (-1)^{1-i} h_{2k+1-i}^{k-1};
   k=1,...,N-1; i=0,...,2k+1

At any k-th step of the recursion you will have an admissible filter
h_n^k_{n=0}^{2k+1}. For example, the initial step (k=0) gives the Haar
filter h_0,h_1; choosing \alpha_1=1/12 after the first step (k=1) you
will have the Daubechies filter h_0,h_1,h_2,h_3.

Inversely, for any admissible filter h_n, n=0,...,2N-1, you can find
N-1 parameter \alpha that recursions above give h_n^{N-1}=h_n.

Similar parametrization can be written for biorthogonal wavelets.

Sincerely yours,
Algirdas

e-mail: algirdas.bastys@maf.vu.lt



Date: Tue, 21 Apr 1998 15:37:26 -0400
From: Paul Costa <pc@mathworks.com>
Subject: #25 Answer: Error in Matlab manual (WD 7.4 #26)

Hello Wavelet Digest authors,

My name is Paul Costa and I work for The MathWorks.  I would like to
answer a question which was posed in the Wavelet Digest, Vol. 07, Nr. 04.

Here is the question:

> Date: Sun, 12 Apr 1998 15:24:03 PKT
> From: "Awais M. Hussain" <awais.hussain@tes-cressoft.com.pk>
> Subject: #26 Question: Error in Matlab wavelet toolbox manual?
> Question:
> The MATLAB wavelet toolbox manual on p.6-13 (1996) defines the DWT 
> of a signal s(n) as:
> C(j,k) = \sum_{n \in Z} s(n) g_{j,k} (n) 
> with 
> g_{j,k}(n) = 2^{-j/2} g( 2^{-j} n - k) 
> where g(n) is the  wavelet filter's impulse response (discrete-
> time). Thus  g_{j,k} are  the dilations of the discrete filter? 
> This does not appear to be the correct representation of the DWT. Can 
> someone explain?
> Awais

Here is the Answer:

There is a typographical error on page 6-13 of the Wavelet Toolbox
User's Guide (March 1996). This error was also present in the MATLAB
5.0 version of this User's Guide (1997) on page 6-14.

The correct equation for g_{j,k}(n) should read:

 g_{j,k}(n) = 2^{j/2} g( 2^{j} n - k) 

Paul Costa
The MathWorks, Inc.
pc@mathworks.com



Date: Thu, 23 Apr 98 09:36:10 -0500
From: "Piotr Wojdyllo" <pwoj@cbr.tpsa.pl>
Subject: #26 Answer: Wavelet Transformation in medical diagnosis (WD 7.4 #17)

   Dear Tang,

   I am a Ph.D. student in Institute of Mathematics, Warsaw
University. In my research led in in Group of Mathematical Logic under
direction of prof. A .Skowron I deal with cases of epilepsy. Our
method combines wavelet analysis of EEG with rough set methods. Our
study is concerned now on detection of posttraumatic cases in a group
suffering epilepsy.

Results are the following: 

1. Known subjects we can recognize with 100% accuracy and unknown
objects we recognize with sensitivity 71% and specifity 75%. Tests
done on half of population.

2. Known subjects we can recognize with 100% accuracy and unknown
objects we can classify with efficiency 86% and 65% depending of
decision class. Tests done on one excluded subject (Of course,
percentage is a mean value for different subjects).

Extended abstract of my work "Wavelets, rough sets and artificial
neural networks in EEG analysis" will appear soon in Wavelet
Digest. The work will appear in Proceedings of Rough Sets and Current
Trends in Computing Conference, Lecture Notes in Computer Science,
Springer Verlag.

By the way, let me know about projects of using wavelets in medical
diagnosis, especially in EEG analysis, you are aware of or engaged in.

Yours sincerely

 Piotr Wojdyllo

Institute of Mathematics
Warsaw University
Banacha 2, 
02-097 Warsaw
POLAND



Date: Tue, 21 Apr 1998 14:24:39 +0900
From: Tang Yipeng <tang@kalsoft.online.sh.cn>
Subject: #27 Answer: Wavelet on medical diagnosis (WD 7.4 #17)

Hi,
You can find many applications of wavelet in biomedical, on the special
issue of 'Annals of Biomedical Engineering' Volume 23 Issue 5
(september/october 1995).
Res. Assist Zeynep Akalin

zakalin@amanos.eee.metu.edu.tr



Date: Tue,  5 May 98 10:45:32 +0100
From: Thomas Harte <thomas.harte@cl.cam.ac.uk>
Subject: #28 Answer: Mathematics of Wavelets (WD 7.4 #16)

Regarding mathematics texts that deal with the requisite mathematics
for wavelets at a level suitable for engineers, by far that best
suggestion that I have had was from Josefina Alvarez of New Mexico
State University.

Josefina suggested:

"Linear Operator Theory in Engineering and Science" (1982)
A. W. Naylor and G. R. Sell, Applied Mathematical
Sciences Series 40, Springer.

The book integrates set theory, topological structure, algebraic
structure, Hilbert spaces &c., and positively bristles with relevant
(and readable) references. It is also written in a tutorial style
without being patronizing yet maintaining succinctness. This is the
book that I have been looking for. At ~600 pages it is short enough to
be unwieldy, yet long enough to cover many topics at a sufficient
level of depth beyond mere regurgitation of definitions as is the case
in many introductary texts on wavelets.

For those who are interested in analysis, I have found the following
to be an enthusiastic and informative exposition which is a useful
read in conjunction with the above:

"An Introduction to Mathematical Analysis"  (1993) J. Lewin
and M. Lewin, International Series in Pure and Applied Mathematics,
McGraw-Hill, 2nd edition.

---
thomas harte @ computer laboratory,  cambridge university;
tph1001@CL.cam.ac.uk; phone: +44 1223 335089; fax: 334678;
http://www.CL.cam.ac.uk/users/tph1001 MIME & NeXT Mail OK.



Date: Thu, 21 May 1998 08:53:49 +0300
From: isar <isar@ee.utt.ro>
Subject: #29 Answer: Ridge and skeleton extraction (WD 7.4 #23) 

The question was:

  >  From: Natalya Hunter Williams <<natalya@pet.hw.ac.uk>
  >  Subject: #23 Question: Ridge and skeleton extraction
  >  
  >  Dear all -
  >  
  >  I am working towards a PhD in Petroleum Engineering, and have
  >  been using wavelets (CWTs) to analyse petrophysical data for
  >  characteristic heterogeneity lengthscales indicative of a
  >  sedimentary control (using Matlab).  I want to apply the ridge
  >  and skeleton method to extract the hierarchical wavelength
  >  information from the data.  I'm looking for some freeware code
  >  which performs this operation.  If you have some ideas about
  >  where I might find this, I'd be very grateful if you could let me
  >  know.

The answer is:

The WaveLab Toolbox of Matlab contains functions to compute the
continuous wavelet transform and to extract ridges and the skeleton.

These functions are CWT, BUildSkelMap and ExtractRidge.

You can download this Toolbox by anonymous FTP to
ftp://playfair.stanford.edu/pub/wavelab. The filnames are:

WaveLab0700.sea.hqx (for Mac)

WaveLab0700.tar.Z   (for UNIX)

wlab0700.zip        (for MS-Windows).

Best whishes.

Alexandre Isar
Technical University Timisoara
isar@ee.utt.ro



Date: Sat, 18 Apr 98 10:16:06 ist
From: "Dr. A. Avudainayagam" <avudai@acer.iitm.ernet.in>
Subject: #30 Question: Rediscovering Haar wavelets & wavelet packets

Recently Hugh Casement (Volume 7, Issue 3) attempted a numerical
experiment which resulted in the rediscovery of Haar wavelets! Here we
want to describe our similar efforts in meddling with Haar wavelets
which appears to give better(!) results when the aim is to make a
vector (representing a signal sampled at equal intervals of time)
sparse.
        We simply swap the bottom n/2 rows and the top n/2 rows in an
nxn Haar matrix and call as H'.
             _                        _
            |  s -s  0  0 . . . . .  . |
            |  0  0  s -s . . . . .  . |
            |  .  .  .  . . . . . .  . |
            |  .  .  .  . . . . . .  . |
            |  0  0  .  . . . . . s -s |
   H'=      |  s  s  0  0 . . . . .  . |      ,  s=1/sqrt(2)
            |  0  0  s  s . . . . .  . |
            |  .  .  .  . . . . . .  . |
            |  .  .  .  . . . . . .  . |
            |  0  0  0  0 . . . . s  s |
	     -                        -
We apply H' in the same way as the Haar matrix on an nx1 vector which
means that we are transforming the details instead of the averages.
For any chosen threshold, H' makes the sampled signal more sparse than
the traditional Haar matrix H.  For example: Consider the signal f(t)
= t squared, 0 <= t < 1 sampled at i/128, 0 <= i <= 127.  Note that
for threshold value = 0.000001, the H'-transformed signal has 16
elements which may be ignored while the conventional H-transformed
signal has no element to be ignored.  Similar results are obtained for
other signals given below.  Threshold value = 0.000001

 _____________________________________________________________________________
|         | number of elements which can be |number of elements which can be  |
|         | ignored in H transformed signal |ignored in H' transformed signal |
|_____________________________________________________________________________|
|         |                                 |                                 |
|t cubed  |        1                        |                 9               |
| sin(t)  |        0                        |                 8               |
| cos(t)  |        0                        |                12               |
| exp(t)  |        0                        |                 8               |
 ------------------------------------------------------------------------------
A similar improvement(!) can also be acheived for compressing matrices (images).
Question: Does this mean H' compresses the signal better than H?



Date: Sun, 19 April 1998 
From: "Raul Machado Garcian" <t96_mam@t.kth.se>
Subject: #31 Question: Wavelets and Vibration Monitory.

Dear Waveleteers,

I am a M.Sc. student of Mechanical/Vehicle Engineering and I am
working on a project which involves analysis of Vibration Monitory by
using wavelets analysis.I have seen few papers about it where formulas
are applied without explaining the mathematical background.

I will be pleased receiving information regarding theses and projects 
where I can find  more mathematics behind vibration monitory with 
wavelets analysis and its applications.

Thanks
Raul Machado Garcia.
e-mail: t96_mam@t.kth.se



Date: Sun, 19 Apr 1998 09:54:06 -0600 (MDT)
From: dkladar@ccinet.ab.ca (Dalibor Kladar)
Subject: #32 Question: Wavelets & Power - Quality or Surge Protection?

Dear Waveleteers,

Has anyone seen any references to wavelets' application in Power
Quality?  i.e., using a Wavelets to analyze various waveforms during
electrical overstresses (from lightning, electrostatic discharges,
switching of reactive loads etc.).

Thanks, 

Dalibor Kladar



Date: Mon, 20 Apr 98 13:25:52 +0100
From: sanjiv-x.sharma@bae.co.uk
Subject: #33 Question: Wavelets & Neural Network characterisation

     Hello,
     
     I am a PhD student researching into Feature Extraction and
     Feature Engineering for use in Neural Network based
     characterisation.  I would like to find out how I may apply
     wavelets to this field.  Being heavily ingrained and conditioned
     by Fourier transforms, I understand the mechanism, but not the
     results of the wavelet transforms.  Is there any literature that
     explains how one should extract information from the wavelet
     transformed domain?
     
     Also I don't fully appreciate the different types of wavelets and
     their applications.  I would be grateful if anyone could either
     direct me to engineering biased literature, or provide some basic
     insight.
     
     I have heard about Wavelet Neural Networks, but have insufficient
     information to apply them.  Again any pointers to literature
     explaining how they work and what they do would also be greatly
     appreciated.
     
	     e-mail:    either  ssharma@euclid.powernet.co.uk
                or      sanjiv-x.sharma@bae.co.uk
     
     Thank you for your assistance
     
     Regards
     Sanjiv



Date: Wed, 22 Apr 98 10:21 +0100
From: r.kerns@t-online.de (Richard Kerns)
Subject: #34 Question: Negative wavelet coefficients

                                                       14 April, 1998

Dear Waveleteers,

     I recently began learning about Wavelets.  I decided to start with
the Haar Wavelet.  I obtained a copy of the software package
"Wavelab", which I used with Matlab.  I first read about the Haar
functions and the Haar series in a book by K. G. Beauchamp called
"Walsh Functions and their applications" published by Academic Press.

     I first tried the function f(x)=1 in the interval from 0 to 1.
Using 64 functional values I obtained 8 for the first Haar coefficient
with the other coefficients being 0.  This differs from the value I
had expected by a factor of 8.  I understand that this factor of 8
arises as the sqrt. of 64, because I used 64 values.  I decided that
in working with wavelab I should change the formula by multiplying by
the norm, which is the sqrt. of the number of functional values used.

     I tried another example in the interval from 0 to 1 with
f(x)=2 for x less than 1/2 and f(x) =0 for x greater than 1/2.  I
calculated that the first 2 Haar coefficients should be 1 and the rest
0 using the formulae in the book by Beauchamp.  When I use 64
functional values to define the function, then I expect using the
changed formula described in the paragraph above, the value 8 for each
of the first two coefficients and 0 for the other coefficients.  Using
Wavelab I then found the first coefficient to be 8 as expected, but
the second coefficient was -8.  I am not able to explain the negative
sign.  After all I had obtained my function by taking the sum of the
first two Haar functions.  I would be grateful for any help you may be
be able to give me.
                            sincerely,
                            Richard Kerns



Date: Wed, 22 Apr 1998 21:42:58 +0200
From: Eric von Lieres <lieres@numerik.math.uni-siegen.de>
Subject: #35 Question: Jackson-type inequality (simultaneous approximation)

Dear Waveletters and Approximators,

For my diploma thesis I need a Jackson-type inequality of the form

||f - P_j f||_W_m_p <= C 2^-((s-m) j) ||f||_W_s_p

where ||.||_W_m_p is the usual Sobolev-Norm on R^d with 1 < p <
infinity and m,s,d integer. P_j denotes the (orthogonal) projector on
the principal shift-invariant approximation space V_j concerning to
the multiresolution-analysis generated by a single scaling-function
Phi.

I use the compactly supported wavelets of I. Daubechies with some
regularity and those of Y. Meyer wich decay faster than every inverse
polynomial.

I know a proof for the case p=2, but it involves fourier-techniques, which
are not that easy to handle in the general case.

Does anyone know results of the requested kind? Hints leading in this
direction are appreciated, too.

Thanks in advance.

Eric von Lieres
Fachbereich Mathematik
Uni-GH Siegen
Hoelderlinstrasse 6
57068 Siegen
<lieres@mail.math.uni-siegen.de>



From: Dorota W <ketdb@eden.tu.kielce.pl>
Subject: #36 Question: Burt-Adelson wavelet

I am looking for an analytical formula for a Burt-Adelson wavelet. It
is implemented in MathCad 6.0+ as a two-dimensional wavelet. The
function is "wave2d". No description is given in MathCad electronic
book called Image processing Function Pack.

Dorota Wiraszka

Kielce University of Technology

Poland



Date: Mon, 27 Apr 1998 11:07:15 -0400
From: "Lijian Zhang" <lzhang@alphorn.sb.fsu.edu>
Subject: #37 Question: Wavelet for surface reconstruction?

Dear waveletters;

I am a fresher in the wavelet. Now I have a set of 3D data. Now I want
use these to reconstruct a 3D surface. Could you tell me can I use
wavelet to solve my problem. I just read one of the book about
wavelet. It seams I can do it, is true?

Thank you in advance for your help.

Cheers --- Lijian

        Dr. Lijian Zhang        
                
        Institute of Molecular Biophysics
        Florida State University
        Tallahassee, FL 32306-4380
        U.S.A
        
        Phone: +1-850-6445606
        Fax:   +1-850-6441406
        Email: lzhang@sb.fsu.edu



Date: Wed, 6 May 1998 15:41:31 +0000
From: "Ralph Quadflieg" <quadflieg@uni-bonn.de>
Subject: #38 Question: Wavelet analysis of biological systems

Hi to All,

I am searching for publications dealing with wavelet analysis of the
response of biological systems to various stimuli.

Are there any expirience in this field ?

Thank you in advance 

Ralph Quadflieg, PharmD
Department of Pharmceutical Technology and Biopharmacy
University of Bonn
D-53121 Bonn
Germany
http://www.pharma.uni-bonn.de
Tel.: ++49 228 735117
Fax.: ++49 228 735268

quadflieg@uni-bonn.de



Date: Fri, 08 May 1998 20:44:09 PDT
From: "Yao Kaisheng" <ksyao@hotmail.com>
Subject: #39 Question: About Shift-Invariant Wavelet

Dear Waveleter,

It is commonly known that shift invariance is very important to signal
processing, and DWT does not have shift invariance property.  I read a
paper, "Multiresolution representations using the auto-correlation
functions of compactly supported wavelet", IEEE trans.  on SP, Vol 41,
1993. The paper proposed a fast algorithm using autocorrelation shell
of Daubechies family and the Mallat algorithm with downsampling to
calculate symmetrical shift invariant wavelet decomposition
coefficients.
  Anybody have done experiments using the proposed algorithm?
  How about the perfomance of the algorithm?

   Yao Kaisheng
   e-mail:  ksyao@mail.cic.tsinghua.edu.cn



Date: Sat, 16 May 1998 13:49:38 +0800
From: Liu Bin <bliu@biosys.net>
Subject: #40 Question: Maximum Absolute Deviation?

Hi everyone,

Would you please tell me something about the Maximum Absolute
Deviation method in WNOISEST.M in MATLAB v5.0 WAVELET TOOLBOX, or
about how to estimate noise level via wavelet transform in order to
use the WDEN.M denoising method?

Thanks.

Please reply to bliu@biosys.net!

[Answer WD 7.6 #20]



Date: Sun, 17 May 98 22:58:24 -0400
From: rickeym <rickeym@hubcap.clemson.edu>
Subject: #41 Question: Wavelet spectral slopes

I'm trying to determine the relationship between Wavelet spectra and 
Fourier spectra.  I've read an article by Valerie Perrier that compares 
Wavelet spectra and Fourier spectra for real wavelets.  However, I use 
the morlet wavlet, which is complex. 

  Using the Morlet wavelet transform on a real signal, Is there a 
relationship between spectral slopes obtained from wavelet spectra and 
spectral slopes obtained from classical Fourier analysis.

Rick McDaniel
Clemson University
rickeym@hubcap.clemson.edu  



Date: Mon, 18 May 1998 10:43:14 +0100 (British Summer Time)
From: adefeyike maryam odutayo <a.m.odutayo@cranfield.ac.uk>
Subject: #42 Question: Introduction to wavelets for non-mathematicians/engineers

Hello,

I am a PhD student with a computer science background.  I need
excellent references on Introduction to wavelets and its application
to image processing.

I have read many journal articles and a couple of textbooks.  All
these have been directed at mathematicans and engineers and have
therefore not been of much assistance to me.

I would appreciate it if any one can refer me to references:
textbooks, journals, websites, etc, with little mathematical detail
detail but deep conceptual information on wavelets and its application
to image processing.

Thank you.

adefeyike.

[Answer WD 7.6 #19]


<webmaster@wavelet.org>