[Home]
[Back Issues]
[Search]
[Vol 1]
[Vol 2]
[Vol 3]
[Vol 4]
[Vol 5]
The Wavelet Digest - Volume 4, Issue 8 (July 28, 1995)
Today's Editor: Wim Sweldens
Interval Research Corporation, Palo Alto
sweldens@interval.com
Today's Topics:
1. Book: Digital Video Processing
2. Preprint: Evaluation of Connection Coefficients
3. Preprint: The Lifting Scheme
4. Preprint: New de-noising papers available from Stanford
5. Preprint: WavBox Software and Related Papers.
6. Preprint: Wavelet technology in a text-to-speech system
7. Preprint: Besov Norms in Terms of the Continuous Wavelet Transform.
8. Preprint: Periodic orthonormal quasi-wavelets
9. Preprint: Wavelet packet transform and seismic signal processing
10. Preprint: B-spline wavelet for multiscale edge detection
11. Meeting: Spring School: 1/f Long Range Dependent Processes
12. Contents: Advances in Computational Mathematics
13. Answer: WD 4.7 (#10) Heisenberg Uncertainty
14. Question: Wavelet and multisampled systems
15. Question: Wavelets for electroencephalographers
16. Question: Wavelet and fractals
17. Question: Concerning a theorem in Chui's book
18. Question: Wavelet Analysis and Broadband Traffic Flows
19. Question: Wavelet transforms and turbulence
Submissions:
E-mail to wavelet@math.sc.edu with "submit" as subject.
Subscriptions:
E-mail to wavelet@math.sc.edu with "subscribe" as subject.
To unsubscribe, e-mail with "unsubscribe" followed by your e-mail
address as subject. To change address, unsubscribe and resubscribe.
Preprints, references, and back issues can be obtained from our
information servers:
Ftp: ftp.math.sc.edu (/pub/wavelet)
Gopher: gopher.math.sc.edu
WWW: http://www.math.sc.edu/~wavelet/
Current number of subscribers: 5092
Calendar of events:
Jul 31-Aug 25: International Summer School, Jyvaskyla, Finland
*Aug 6: SIGGRAPH95: Wavelet course, LA
*Aug 30-31: UK Symposium on Time-Frequency/Scale Methods, Warwick
Sep 11-15: UCLA short course on "Wavelet Transform Applications"
Sep 17-21: ASME Wavelets in Vibrations and Acoustics, Boston
Sep 20-23: Advances In Biomedical Signal And Image Proc, Montreal
*Nov 27-Dec 1: Acoustical Society of America, St. Louis
Dec 10-13: Neural Networks and Signal Processing, Nanjing, China
-- 1996 --
Jan 28-Feb 2: Digital Video Compression, San Jose, CA
Jan 27-Feb 2: Still-Image Compression, Photonics West, San Jose
*Apr 29-May 3: 1/f Long Range Dependent Processes, Les Houces, Fr.
From: A. Murat Tekalp <tekalp@ee.rochester.edu>
Subject: Book: Digital Video Processing
Digital Video Processing, by A. Murat Tekalp
Prentice Hall, ISBN 0-13-190075-7, available August 1995
This 550 page textbook covers various aspects of digital video processing
technology in a tutorial style suitable for classroom teaching or self-study.
CONTENTS
PART 1: Representation of Digital Video
Chapter 1 Basics of Video
Chapter 2 Time-Varying Image Formation Model
Chapter 3 Spatio-Temporal Sampling
Chapter 4 Sampling Structure Conversion
PART 2: Two-Dimensional Motion Estimation
Chapter 5 Optical Flow Methods
Chapter 6 Block-Based Methods
Chapter 7 Pel-Recursive Methods
Chapter 8 Bayesian Methods
PART 3: Three-Dimensional Motion Estimation and Segmentation
Chapter 9 Methods Using Point Correspondences
Chapter 10 Optical Flow and Direct Methods
Chapter 11 Motion Segmentation
Chapter 12 Stereo and Motion Tracking
PART 4: Video Filtering
Chapter 13 Motion Compensated Filtering
Chapter 14 Noise Filtering
Chapter 15 Restoration
Chapter 16 Standards Conversion
Chapter 17 Superresolution
PART 5: Still Image Compression
Chapter 18 Lossless Compression
Chapter 19 DPCM and Transform Coding
Chapter 20 Still Image Compression Standards
Chapter 21 Vector Qunatization, Subband Coding and Other Methods
PART 6: Video Compression
Chapter 22 Interframe Compression Methods
Chapter 23 Video Compression Standards
Chapter 24 Model-Based Coding
Chapter 25 Digital Video Systems
APPENDICES:
Appendix A: Markov and Gibbs Random Fields
Appendix B: Basics of Segmentation
Appendix C: Kalman Filtering
For further info, please contact:
A. Murat Tekalp tekalp@ee.rochester.edu
Karen Gettman, Prentice-Hall kgettman@cup.portal.com
Ann Sullivan, Prentice-Hall ann@prenhall.com
WWW: http://www.ceas.rochester.edu:8080/ee/faculty/tekalp.html
From: Eric Tenenbaum <tene@BBN.COM>
Subject: Preprint: Evaluation of Connection Coefficients
Hello fellow Waveleteers
I have been made aware (no pun intended) that numerous people
have expressed an interest in receiving a copy of the paper,
"The Evaluation of Connection Coefficients of Compactly Support Wavelets"
by A. Latto, H. L. Resnikoff and E. Tenenbaum.
I would be glad to mail a copy to anyone that makes a request. Feel
free to email me your street address and I will do my best to respond
in a timely fashion. My email address is tene@bbn.com.
Eric Tenenbaum
From: Wim Sweldens <sweldens@interval.com>
Subject: Preprint: The Lifting Scheme
Title: The Lifting Scheme: A Construction of Second Generation Wavelets
Author: Wim Sweldens
Abstract:
We present the lifting scheme, a simple construction of second
generation wavelets, wavelets which are not necessarily translates and
dilates of one fixed function. Such wavelets can be adapted to
intervals, domains, surfaces, weights, and irregular samples. We show
how the lifting scheme leads to a faster, in-place calculation of the
wavelet transform. Several examples are included.
Report 1995:05, Industrial Mathematics Initiative, Department of
Mathematics, University of South Carolina.
Anonymous ftp to ftp.math.sc.edu, file /pub/imi_95/imi95_6.ps or imi95_6.ps.gz.
From: David Donoho <donoho@playfair.Stanford.EDU>
Subject: Preprint: New de-noising papers available from Stanford
De-Noising Papers available by FTP / WWW
The following preprints are available electronically:
1. WaveLab and Reproducible Research, by J. Buckheit and D. Donoho
2. Atomic Decomposition by Basis Pursuit, by S. Chen and D. Donoho
3. Translation-Invariant De-Noising, by R. Coifman and D. Donoho
4. Wavelet threshold estimators for data with correlated noise
by I. Johnstone and B. Silverman
5. Ideal Denoising in an orthonormal basis chosen from a library of bases
by D. Donoho and I. Johnstone
The papers are available in compressed postscript format (.ps.Z).
Instructions for WWW Access:
* Open the URL http://playfair.stanford.edu/subjects/wavelets.html
* Click on the name of the paper you desire.
Instructions for FTP Access:
The FTP addresses of this papers are
1. ftp://playfair.stanford.edu/pub/buckheit/wavelab.ps.Z
2. ftp://playfair.stanford.edu/pub/chen_s/BasisPursuit.ps.Z
3. ftp://playfair.stanford.edu/pub/donoho/TIDeNoise.ps.Z
4. ftp://playfair.stanford.edu/pub/johnstone/correlated.ps.Z
5. ftp://playfair.stanford.edu/pub/donoho/idealbasis.ps.Z
To get #1, for example, do the following
anonymous ftp to playfair.stanford.edu
cd pub/buckheit
binary
get wavelab.ps.Z
Instructions for the other papers are similar.
Instructions for printing:
UNIX users should remember that in many systems "lpr file.ps"
will not work properly if file.ps is greater than 1.25MB in length.
It is probably necessary to find a workaround for large postscript files
on your UNIX. On many UNIX systems, "lpr -s file.ps" will work.
Questions: donoho@playfair.stanford.edu
Abstracts:
\title{WaveLab and Really Reproducible Research}
\author{Jon Buckheit \& David L. Donoho \\
Statistics Dept., Stanford University}
\begin{abstract} \WaveLab is a library of \Matlab routines for wavelet
analysis, wavelet-packet analysis, cosine-packet analysis and matching
pursuit. The library is available free of charge over the Internet.
Versions are provided for Macintosh, UNIX and Windows machines.
\WaveLab makes available, in one package, all the code to reproduce all
the figures in our published wavelet articles. The interested reader can
inspect the source code to see exactly what algorithms were used, how
parameters were set in producing our figures, and can then modify the
source to produce variations on our results. \WaveLab has been developed,
in part, because of exhortations by Jon Claerbout of Stanford that
computational scientists should engage in ``really reproducible''
research. \end{abstract}
\title{Atomic Decomposition by Basis Pursuit}
\author{Shaobing Chen \& David L. Donoho \\
Statistics Dept., Stanford University}
\begin{abstract} The Time-Frequency and Time-Scale communities have
recently developed a large number of overcomplete waveform dictionaries --
stationary wavelets, wavelet packets, cosine packets, chirplets, and
warplets, to name a few. Decomposition into overcomplete systems is not
unique, and several methods for decomposition have been proposed --
including the Method of Frames, Matching Pursuit, and, for special
dictionaries, the Best Orthogonal Basis.
Basis Pursuit is a principle for decomposing a signal into an ``optimal''
superposition of dictionary elements -- where optimal means having the
smallest $l^1$ norm of coefficients among all such decompositions. The
principle exhibits several advantages over the Method of Frames, Matching
Pursuit and Best Ortho Basis, including better sparsity, and
super-resolution. BP has interesting relations to ideas in areas as
diverse as ill-posed problems, in abstract harmonic analysis, total
variation de-noising, and multi-scale edge de-noising.
Basis Pursuit in highly overcomplete dictionaries leads to large-scale
optimization problems. With signals of length 8192 and the wavelet packet
dictionary, one gets an equivalent linear program of size 8192 by 212,992.
Such problems can be attacked successfully because of recent advances in
linear programming by interior point methods. We obtain reasonable success
with a primal-dual logarithmic potential method and conjugate gradient
solver. \end{abstract}
{\bf Key Words and Phrases.} Overcomplete signal representation,
Time-Frequency Analysis, Time-Scale Analysis, $\ell^1$ norm optimization,
Matching Pursuit, Wavelets. Wavelet Packets, Cosine Packets, Interior
point methods for linear programming, Total Variation de-noising,
multi-scale edges.
\title{Translation-Invariant De-Noising}
\author{David Donoho\\Department of Statistics\\Stanford University
\and
R.R. Coifman\\Mathematics\\Yale University}
\begin{abstract} We discuss a method for making wavelet thresholding
de-noising translation invariant. It is based on averaging the results of
processing data arising in all possible circulant shifts of the original
data. There are fast algorithms to accomplish this in order $n \log n$
time. The method is related to simple thresholding in Mallat's
undecimated wavelet transform. The method has a number of performance
benefits; for example when used with the Haar wavelet it suppresses Gibbs
phenomena in expectation. Similar ideas make sense with cosine packets,
wavelets packets and other orthogonal bases. \end{abstract}
\title{Wavelet threshold estimators for data with correlated noise}
\author{Iain M. Johnstone\\Department of Statistics\\Stanford
University\\Stanford CA 94305\\U.S.A. \and
Bernard W. Silverman\\School of
Mathematics\\Bristol University\\Bristol BS8 1TW %
\begin{abstract}
Wavelet threshold estimators for data with stationary correlated noise are
constructed by the following prescription. First, form the discrete
wavelet transform of the data points. Next, apply a {\em level-dependent}
soft threshold to the individual coefficients, allowing the thresholds to
depend on the level in the wavelet transform. Finally, transform back to
obtain the estimate in the original domain. The threshold used at level
$j$ is $s_j \sqrt{2 \log n}$, where $s_j$ is the standard deviation of the
coefficients at that level, and $n$ is the overall sample size. The
minimax properties of the estimators are investigated by considering a
general problem in multivariate normal decision theory, concerned with the
estimation of the mean vector of a general multivariate normal
distribution subject to squared error loss. An ideal risk is obtained by
the use of an `oracle' that provides the optimum diagonal projection
estimate. This `benchmark' risk can be considered in its own right as a
measure of the sparseness of the signal relative to the noise process, and
in the wavelet context it can be considered as the risk obtained by ideal
spatial adaptivity. It is shown that the level-dependent threshold
estimator performs well relative to the benchmark risk, and that its
minimax behaviour cannot be improved upon in order of magnitude by any
other estimator. \end{abstract}
{\it Key Words and Phrases:} Decision
theory, Level-dependent thresholding, Minimax estimation, Nonlinear
estimators, Nonparametric regression, Oracle inequality, Wavelet
transform.
\title{Ideal Denoising in an orthonormal basis chosen from a library of
bases}
\author{David L. Donoho \& Iain M. Johnstone\\
Department of Statistics\\ Stanford University}
\begin{abstract} Suppose we have observations $y_i = s_i + z_i$,
$i=1,...,n$, where $(s_i)$ is signal and $(z_i)$ is i.i.d. Gaussian white
noise. Suppose we have available a library $\cL$ of orthogonal bases, such
as the Wavelet Packet bases or the Cosine Packet bases of Coifman and
Meyer. We wish to select, adaptively based on the noisy data $(y_i)$, a
basis in which best to recover the signal (``de-noising''). Let $M_n$ be
the total number of distinct vectors occcuring among all bases in the
library and let $t_n = \sqrt{2 \log(M_n)}$. (For wavelet packets, $M_n = n
\log_2(n)$.)
Let $y[\cB]$ denote the original data $y$ transformed into the Basis
$\cB$. Choose $\lambda > 8$ and set $\Lambda_n = (\lambda \cdot (1 +
t_n))^2$. Define the entropy functional
\[
\cE_\lambda(y,\cB) = \sum_i \min(y_i^2[\cB], \Lambda_n^2 ) .
\]
Let $\hat{\cB}$ be the best orthogonal basis according to this entropy:
\[
\hat{\cB} = \argmin_{\cB \in \cL} \cE_\lambda(y,\cB) .
\]
Define the hard-threshold nonlinearity $\eta_t(y) = y 1_{\{|y| > t\}}$.
In the empirical best basis, apply hard-thresholding with threshold $t =
\sqrt{\Lambda_n}$:
\[
\hat{s}_i^*[\hat{\cB}] = \eta_{\sqrt{\Lambda_n}} (y_i[\hat{\cB}]) .
\]
{\it Theorem:} {\sl With probability exceeding $\pi_n = 1-e/M_n$,
\[
\|\hat{s}^* - s\|_2^2 \leq (1-8/\lambda)^{-1} \cdot \Lambda_n \cdot
\min_{\cB \in \cL} E \| \hat{s}_{\cB} - s \|_2^2 .
\]
Here the minimum is over all ideal procedures working in all bases of the
library, i.e. in basis $\cB$,
$\hat{s}_{\cB}$ is just $y_i[{\cB}] 1_{\{|s_i[\cB]| > 1\}}$.}
In short, the basis-adaptive estimator achieves a loss within a
logarithmic factor of the ideal risk which would be achievable if one had
available an oracle which would supply perfect information about the ideal
basis in which to de-noise, and also about which coordinates were large or
small.
The result extends in obvious ways to more general orthogonal basis
libraries, basically to any libraries constructed from an at-most
polynomially-growing number of coefficient functionals. Parallel results
can be developed for closely related entropies. \end{abstract}
{\bf Key Words.} Wavelet Packets, Cosine Packets, weak-$\ell^p$ spaces.
Adaptive Basis Selection. Oracles for adaptation. Thresholding of Wavelet
Coefficients.
From: taswell@wavbox.com (Carl Taswell)
Subject: Preprint: WavBox Software and Related Papers.
WavBox Software and Related Papers.
The following manuscript has been submitted for publication; it
is available from the author.
Satisficing Search Algorithms for Selecting Near-Best Bases
in Adaptive Tree-Structured Wavelet Transforms
Satisficing search algorithms have been proposed for adaptively
selecting near-best basis decompositions in redundant tree-structured
wavelet transforms. Any of a variety of additive or non-additive
information cost functions can be used as the decision criterion for
comparing and selecting nodes when searching through the tree. Thus,
the algorithms are applicable to transforms generated by any kind of
wavelet whether orthogonal, biorthogonal, or non-orthogonal. These
satisficing search algorithms implement sub-optimizing rather than
optimizing principles, and acquire the important advantage of reduced
computational complexity with significant savings in memory, flops, and
time. Despite the sub-optimal approach, top-down tree search algorithms
with additive or non-additive costs that yield near-best bases are as
good as bottom-up tree search algorithms with additive costs that yield
best bases. Experimental results comparing the various information cost
functions and basis selection methods are demonstrated for both data
compression of real speech and time-frequency analysis of artificial
transients.
The following abstract has been submitted for presentation at the
1995 MATLAB Conference to be held October 16-18 in Cambridge, MA;
further information about WavBox software is available from the author.
WavBox 4.2 -- A Software Toolbox for Wavelet Transforms and Adaptive
Wavelet Packet Decompositions with New Satisficing Search Algorithms
WavBox provides both a function library and a computing environment for
wavelet transforms and adaptive wavelet packet decompositions. WavBox
contains a collection of these transforms, decompositions, and related
functions that perform multiresolution analyses of 1-D multichannel
signals and 2-D images. The older version 4.1 includes overscaled
pyramid transforms, discrete wavelet transforms, and adaptive wavelet
and cosine packet decompositions by best basis and matching pursuit as
described by Mallat, Coifman, Wickerhauser, and other authors, as well
as Donoho and Johnstone's wavelet shrinkage denoising methods. The new
version 4.2 also implements Taswell's satisficing search algorithms for
the selection of near-best basis decompositions with either additive or
non-additive information costs. Various choices of filter classes
(orthogonal, biorthogonal, etc), filter families (Daubechies, Vetterli,
etc), and convolution versions (interval, circular, extended, etc) exist
for each transform and decomposition. The software has been designed for
efficient automated computation, interactive exploratory data analysis,
and pedagogy. Essential features of the design include: perfect
reconstruction for multiresolution decomposition of data of arbitrary
size not restricted to powers of 2; both command line and graphical user
interfaces with a comprehensive set of plots and visual displays; an object
property expert system with artificial intelligence for configuring valid
property combinations; heirarchical modules and switch-driven function
suites; vector-filter and matrix-operator implementations of convolutions;
extensibility for the inclusion of other wavelet filters, convolution
versions, and transforms; optional arguments with built-in defaults for
most m-files; and extensive on-line help and self-running tutorial demos.
Carl Taswell is the author of both WavBox software and the above related
papers.
He can be reached at "taswell@sccm.stanford.edu" or at "taswell@wavbox.com".
From: mei@trlvm.vnet.ibm.com
Subject: Preprint: Wavelet technology in a text-to-speech system
Paper on wavelet technology used in a text-to-speech (TTS) system
and description of and purchasing info about the TTS system.
IBM Tokyo Research Laboratory, Research Report RT0110, June 1995
ProTALKER: a Japanese text-to-speech system for personal computers
(Report is in English.)
Takashi SAITO, Masaharu SAKAMOTO, Yasuhide HASHIMOTO,
Mei KOBAYASHI, Masafumi NISHIMURA, Kazuhiro SUZUKI
IBM Japan, Ltd., Tokyo Research Laboratory,
1623-14, Shimotsuruma, Yamato-shi, Kanagawa-ken 242 Japan.
Abstract:
This paper describes the development of ProTALKER, a Japanese
text-to-speech (TTS) system for personal computers. The quality of
synthesized speech is one of the most important features of any TTS
system. Synthesis methods which are based on manipulation of the speech
signal spectrum (e.g. linear predictive coding synthesis and formant
synthesis) produce comprehensible but unnatural sounding output. The
lack of naturalness commonly associated with these methods results from the
use of oversimplified speech models, small synthesis unit inventories,
and poor handling of text parsing for prosody control. We developed four
new technologies to overcome these difficulties and improve the quality
of output from TTS systems: accurate pitch mark determination by wavelet
analysis, speech waveform generation using a modified time domain pitch
synchronous overlap-add method, speech synthesis unit selection using
a context dependent clustering method, and efficient prosody control
using a 3-phrase parser.
For further info about the report, pls. contact:
Mei Kobayashi (LAB-s73)
IBM Tokyo Research Laboratory
1623-14 Shimotsuruma, Yamato-shi, Kanagawa-ken 242 Japan
tel: 81+462-73-4934, FAX: 81+462-73-7428
e-mail: MEI@TRLVM.VNET.IBM.COM
(during Aug 6-24 only, pls. e-mail: kobayash@math.berkeley.edu)
ProTALKER, version 1.0, a product of IBM Japan, Ltd., reads aloud
Japanese kana-kanji text. It can be installed on any PC with audio
capability and Japanese MS-Windows 3.1, such as the IBM ThinkPad. Users
may choose between 4 levels of speech quality; 2.6MB to 18MB are needed
for dictionary storage, with higher quality speech output requiring
more disk space for a better TTS dictionary. Options include: selection
of gender and facial image of speaker,and volume/loudness, speaking
rate, and intonation control. A cartoon face appears in a smaller
window with two columns of buttons, which look and function like those
of a tape recorder, i.e. start, stop, rewind, and fast forward, plus a
help button, which labelled by a question mark symbol. Any length text
file can be read by ProTALKER, however, only 32MB at a time can be
placed in the application buffer. Special options of ProTALKER include:
converting the text file to a separate waveform file after reading,
appending specialized or technical words to the TTS dictionary, and
invoking a forum utility feature to read aloud contents of user
specified forum file, such as e-mail or net news.
list price: 9,800 yen
for purchasing info., pls. contact:
Nakamura (0462)-73-5932 or Takao (0462)-73-2512
IBM Japan, Ltd., Yamato Development Laboratory,
1623-14, Shimotsuruma, Yamato-shi, Kanagawa-ken 242 Japan.
From: perrier@lmd.ens.fr (PERRIER Valerie)
Subject: Preprint: Besov Norms in Terms of the Continuous Wavelet Transform.
BESOV NORMS IN TERMS OF THE CONTINUOUS WAVELET TRANSFORM
APPLICATION TO STRUCTURE FUNCTIONS
by Valerie Perrier & Claude Basdevant
perrier@lmd.ens.fr
The continuous wavelet transform is extended to $L^p$ spaces and an
inversion formula is demonstrated. From this the Besov spaces can be
characterized by the behaviour at small scales of the wavelet
coefficients. These results apply to the measurement of structure
functions.
Mathematical Models and Methods in Applied Sciences
To appear (1996)
From: chen@math03.math.ac.cn (Chen Han Lin)
Subject: Preprint: Periodic orthonormal quasi-wavelets
Abstract
In this paper we present a new method in constructing the periodic
orthonormal quasi-wavelet in $\bo L_2(I)$ with any desired regularity
which differs from Meyer's and Danbechies' periodic wavelet [Me], [Dau].
The interesting thing is that the decomposition and reconstruction
formulas are very simple, each of them involves only two terms. As for
the error estimate, the convergence order is the best in comparism with
the results obtained by other authors. In the second part of this paper,
we give am application of periodic quasi-wavelet to solve a kind of
integral equations. By using the projection method, we obtain a system
of equations, it is important to point out that the coefficients matrix
is diagonal and therefore, is easy to solve.
From: Wang Yuping <ylcai@xjtu.edu.cn>
Subject: Preprint: Wavelet packet transform and seismic signal processing
Title: Application of wavelet packet transform to seismic signal processing
Authors: Yu Ping Wang, Yuan Long Cai., Zhong Xing Geng and Rui Feng
Abstract:
One of the important tasks in seismic signal processing is the
accurate measurement of the arrival time of specified seismic
phases or the dispersion of surface waves.However the traditional
time -frequency analysis such as the windowed Fourier transform(WFT)
or Wigner ville distribution(WVD) can not detect them with a high
precision with a short time sequence.In this paper we will show
that wavelet packet transform is a very efficient tool for the instaneous
spectrucm analysis which is well adapted to the representation and
processing of seismic signal.Moreover,the authors carray out an
investigation on the wavelet packet decomposition algorithm,the choice
of best wavelet packet basis,and the representaion of the signal in
the phase plane,etc.We have found Lemarrie-Battle quadrature filters
result in better consequency than Daubechies filters for its symmetry.
Some examples are given to illustrate the processing of the simulated
and realistic data.In particular,the processing of an acoustic cross-borehole
tomography data obtained in oil field demonstrates its high precision in
determining the arrival time of acoustic wave with high frequency in a noisy
background.A variety of approaches such as phase tracing ,stacking and
convolution have difficulties with this problem.Our research shows that
wavelet packet transform provides a new approach to seismic signal
processing.
For details,please contact with the authors.
Wang Yuping or Cai Yuanlong
Image Processing Center
School of Electronic & Information Engineering
Xi'an Jiaotong University,Shaanxi,710049
P.R.C.
Tel:86-29-3268676
E_mail:ylcai@xjtu.edu.cn
From: Wang Yuping <ylcai@xjtu.edu.cn>
Subject: Preprint: B-spline wavelet for multiscale edge detection
Title: Construction and properties of B-spline wavelet
for multiscale edge detection
Authors: Yu Ping Wang, Yuan Long Cai.,
Part of the paper will appear at 1995 IEEE international conferences
on image processing. For details,refer to <<Science in China>>(in
English) Series A,vol.38,no.4,pp.499-512.
Abstract:
B-spline plays a very important role in wavelet theory.Starting from
B-splne, one can construct orthonomal spline wavelet and
non-orthogonal wavelet. One can derive fast implementaion of
continous wavelet transform.etc.All seems in that polynomial spline
function space as the basis of translated and dilated B-spline
constitutes a multiresolution or multiscale approximation of L^2(R)
space. Many authors such as M.Unser,C.K.Chui,J.Z.Wang have made
contribution to this subject. In this paper we aim to use B-spline to
construct wavelet for multiscale edge detection inclulding both the
zero-crossing detection and the maximum detection. Such a transform
is called dyadic wavelet transform originally proposed by S.Mallt.
The edges of a signal can be efficienctly represented and detected
through its zero-crossing or maximum.For B-spline of any order n,the
fast algorithm for decomposition and reconstruction have been
derived.Also the impulse and frequency responses of the corresponding
decomposition and reconstruction filters are given explicitly.We will
show Mallat's wavelet for maximum detection is just one example of the
wavelet designed by us.In terms of time/frequency localization we will
show cubic B-spline is nearly optimal for most applictions.One has the
freedom using the proposed B-spline wavelet to do many applications
such as multiedge-based compression,stereo matching,denosing,conner
detection.etc. To our knowledge,Gaussian functon or its molulation as
Gabor function has occupied a great position in computer vision and
image processing for a long time. Most scale-space approaches are
based on these functions.For its good properties and similarity to
Gaussian function,we believe that B-spline combinaitng wavelet theory
will have great potentiality in computer vision.
Comments are welcome,please contact with the authors.
Wang Yuping or Cai Yuanlong
Image Processing Center
School of Electronic & Information Engineering
Xi'an Jiaotong University,Shaanxi,710049
P.R.C.
Tel:86-29-3268676
E_mail:ylcai@xjtu.edu.cn
[Answer]
From: flandrin@physique.ens-lyon.fr
Subject: Meeting: Spring School: 1/f Long Range Dependent Processes
Spring School --- Preliminary Announcement
1/F LONG-RANGE DEPENDENT PROCESSES,
FROM MODELS TO APPLICATIONS
April 29 -- May 3, 1996
Les Houches, France
Stochastic processes with long-range dependencies are naturally
observed in areas such as hydrology, geophysics or finance, as well as
in road traffic, telecommunication networks or biology. Those
processes, which evidence a 1/f spectral behavior over wide ranges of
frequencies and self-similarity properties, give rise to challenging
problems from the point of view of both analysis and modeling. The aim
of this Spring School is to address such issues from the many
perspectives offered by recent developments in statistics, signal
theory, wavelets and fractal analysis. A series of lectures will be
delivered by a dozen of invited speakers and the number of
participants will be limited to fifty. Registration fees will be of
about FF 2800 --- with possibly partial support --- for the complete
Session, including full accommodation (lodging and meals) at the
Physics Center of Les Houches, a resort village in the French Alps.
Session Directors:
P. Flandrin (ENS Lyon)
J. Levy-Vehel (INRIA Rocquencourt)
Organizing Committee:
P. Abry (ENS Lyon)
P.-O. Amblard (CEPHAG-ENSIEG Grenoble)
E. Moulines (Telecom Paris)
Preliminary list of invited speakers:
A. Arneodo (Bordeaux)
A. Benveniste (Rennes)
J. Beran (Konstanz)
B. Castaing (Grenoble)
Y. Meyer (Paris)
G. Oppenheim (Paris)
M. Shlesinger* (Arlington)
M. Taqqu* (Boston)
M. Teich (New York)
G. Wornell* (Boston)
* to be confirmed
For further information, send an E-mail message to:
lrd96@physique.ens-lyon.fr
From: publish@baltzer.nl (Baltzer Science Publishers)
Subject: Contents: Advances in Computational Mathematics
Advances in Computational Mathematics, Volume 4, No. 1-2, 1995, ISSN 1019
7168 Editors-in-Chief: John C. Mason & Charles A. Micchelli
Contents Volume 4, no. 1-2, 1995
Special issue - MULTISCALE TECHNIQUES. Editor: Wolfgang Dahmen
Preface
pp. 1-26, F. Keinert, Numerical stability of biorthogonal wavelet transforms
pp. 27-50, S. Zeng, C. Vuik and P. Wesseling, Numerical solution of the
incompressible Navier--Stokes equations by Krylov subspace and multigrid
methods
pp. 51-82, K. Urban, On divergence-free wavelets
pp. 83-110, P.W. Hemker, Sparse-grid finite-volume multigrid for 3D-problems
pp. 111-126, S.C. Brenner, A two-level additive Schwarz preconditioner for
the stationary Stokes equations
pp. 127-144, B. Koren and B. van Leer, Analysis of preconditioning and
multigrid for Euler flows with low-subsonic regions
pp. 145-170, A. Kunoth, Multilevel preconditioning -- Appending boundary
conditions by Lagrange multipliers
pp. 171-206, M. Griebel and P. Oswald, Tensor product type subspace
splittings and multilevel iterative methods for anisotropic problems
Submissions of articles and proposals for special issues are to be
addressed to the Editors-in-Chief:
John C. Mason
School of Computing and Mathematics, University of Huddersfield,
Queensgate, Huddersfield, HD1 3DH, United Kingdom
E-mail: j.c.mason@hud.ac.uk
or
Charles A. Micchelli
Mathematical Sciences Department
IBM Research Center
P.O. Box 218, Yorktown Heights, NY 10598, USA E-mail: cam@yktvmz.bitnet
From: maury@tellabs.com (Maurice Givens)
Subject: Answer: Heisenberg Uncertainty (WD 4.7 #10)
Hichri Haikel asked for further explanation of the Heisenberg
uncertainty with regard to the Gabor's windowed Fourier transform.
I hope this helps.
The analysis is performed in both time and frequency.
First, as the window is moved along the signal, the Foruier transform
that is performed within the window will record the change in the
frequency content of the signal as a function of time.
The windowed Fourier transform can also be perceived as
a series of filter banks. Each filter bank records the amplitude
and phase of the content of the signal in its bandpass, as a function
of time.
The sliding window concept implies that a short window is needed for
good resolution in time. However, if the concept of filter banks is
used to describe the windowed Fourier transform, then the filter banks
should have narrow bandwidths for good frequency resolution. In other
words, to get good time resolution (dt) you need a short window. For
good frequency resolution (df) you need a narrow filter bank. These
are opposing concepts (short window implies wide filter). Consequently,
to use the windowed Fourier transform you have to make a trade-off
using the Heisenberg uncertainty
1
dt*df > ------
4*pi
One advantage the Gaussian window has is that it satifies this the
lower bound of this inequality.
What the continuous wavelet transform uses is a constant Q filter
bank. Therefore, the uncertainty becomes
df = c
----
f
Now, with this constant relationship, the analyzing filter can achieve
arbitrarily good time resolution at high frequencies, and arbitrarily
good frequency resolution at low frequencies. It is fortunate that most
real-world signals consist of short-duration high-frequency components
and long-duration low-frequency components.
From: hpuyosa@sait01.plc.um.es (Hector David Puyosa Pi\qa)
Subject: Question: Wavelet and multisampled systems
Hello,
I am studying wavelets and its application to multisampled
system. The idea is take advantage of multiresolution decomposition
for analyse nonstationary system, at multiple times-scales.
I am looking for information about related papers to
control and identification using wavelets.
I hope I will get any input from you,
Thanks
Hector David Puyosa Piqa
Email: hpuyosa@plc.um.es
From: Israel_Yaar@brown.edu (Israel Yaar, M.D.)
Subject: Question: Wavelets for electroencephalographers
Do you know of resources -- texts, software ect -- that will help me
as a physician/Electroencephalographer/electromyographer to understand
wavelets and their applications in clinical-signal-analysis without
having the needed math capabilities?
From: hajimow@Engn.Uwindsor.Ca (Hajimowlana Sayed H)
Subject: Question: Wavelet and fractals
Hi; I want to use fractals (or wavelet) for recognition of some
background images (like cloth or any kind of other materials like
patterned papers) to extract any kind of possible defects in
production lines.I should be able to identify the defects and their
location in order to stop the production line.First I wantto extract
the background pattern and then identify the defects.I would
appreciate if anybody inform me of any available sources and
researches done and whether it is possible to accomplish this by
fractals or not.
Yours Sincerely
Hossain
From: "Norman C. Corbett" <nccorbet@jeeves.uwaterloo.ca>
Subject: Question: Concerning a theorem in Chui's book
Hi,
I have a question regarding a theorem quoted in the book An Introduction
to Wavelets, by Charles K. Chui. On page 140 the author quotes a theorem
(Theorem 5.15), which he attributes to Wiener. However, the author does
not reference this theorem. I was wondering if anyone was aware of any
references which go explain the details of this theorem. In particular, I
am interested in finding out whether or not such Laurent series have
non-trivial annuluses of convergence.
Thanks for your time
Norm Corbett
Department of Applied Mathematics
University of Waterloo
Waterloo, Ontario
Canada, N2L-3G1
email: nccorbett@jeeves.uwaterloo.ca
phone: (519) 885-1211 extension 5289
From: John Amenyo <jta@ans.net>
Subject: Question: Wavelet Analysis and Broadband Traffic Flows
Is anyone aware of any published work on Wavelet analysis
of packet/cell switching traffic flows?
For example, collecting LAN/WAN traffic at an end/intermediate
node and performing wavelet analysis on the resulting time series
(for any purpose).
From: Rickey McDaniel <rickeym@hubcap.clemson.edu>
Subject: Question: Wavelet transforms and turbulence
I have 2 questions:
1. Can the wavelet transform be used on a time series
to analyze for chaos or turbulence.
2. Can the FWT (fast wavelet transform) be used on data
that are nonequally spaced.
<webmaster@wavelet.org>