Wavelet Analysis Basic Concepts and Applications 1St Edition Sabrine Arfaoui Online Ebook Texxtbook Full Chapter PDF

Wavelet Analysis: Basic Concepts and
Applications 1st Edition Sabrine

Arfaoui
Visit to download the full and correct content document:
https://ebookmeta.com/product/wavelet-analysis-basic-concepts-and-applications-1st-
edition-sabrine-arfaoui/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Environmental Biotechnology Basic Concepts and

Applications 1st Edition Viswanath Buddolla
https://ebookmeta.com/product/environmental-biotechnology-basic-
concepts-and-applications-1st-edition-viswanath-buddolla/
Introductory Immunology: Basic Concepts for

Interdisciplinary Applications. 3rd Edition Actor
https://ebookmeta.com/product/introductory-immunology-basic-
concepts-for-interdisciplinary-applications-3rd-edition-actor/
Pharmacokinetic and Pharmacodynamic Data Analysis

Concepts and Applications 5th Edition Johan Gabrielsson
https://ebookmeta.com/product/pharmacokinetic-and-
pharmacodynamic-data-analysis-concepts-and-applications-5th-
edition-johan-gabrielsson/
Big Data Analysis for Green Computing: Concepts and

Applications 1st Edition Rohit Sharma
https://ebookmeta.com/product/big-data-analysis-for-green-
computing-concepts-and-applications-1st-edition-rohit-sharma/
Wavelet Numerical Method and Its Applications in
Nonlinear Problems You He Zhou
https://ebookmeta.com/product/wavelet-numerical-method-and-its-
applications-in-nonlinear-problems-you-he-zhou/
Primary Mathematics 3A Hoerst
https://ebookmeta.com/product/primary-mathematics-3a-hoerst/
Heterophase Polymerization Basic Concepts and

Principles 1st Edition Hugo Hernandez
https://ebookmeta.com/product/heterophase-polymerization-basic-
concepts-and-principles-1st-edition-hugo-hernandez/
Hydrochemistry Basic Concepts and Exercises 2nd Edition

Eckhard Worch
https://ebookmeta.com/product/hydrochemistry-basic-concepts-and-
exercises-2nd-edition-eckhard-worch/
Rowland and Tozer s Clinical Pharmacokinetics and

Pharmacodynamics Concepts and Applications Concepts and
Applications 5th Edition Derendorf
https://ebookmeta.com/product/rowland-and-tozer-s-clinical-
pharmacokinetics-and-pharmacodynamics-concepts-and-applications-
concepts-and-applications-5th-edition-derendorf/
Wavelet Analysis
Wavelet Analysis
Basic Concepts and Applications
Sabrine Arfaoui
University of Monastir, Tunisia
University of Tabuk, Saudi Arabia
Anouar Ben Mabrouk

University of Kairouan, Tunisia
University of Monastir, Tunisia
University of Tabuk, Saudi Arabia
Carlo Cattani
University of Tuscia, Italy
First edition published 2021
by CRC Press
6000 Broken Sound Parkway NW, Suite 300, Boca Raton, FL 33487-2742
and by CRC Press

2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
© 2021 Sabrine Arfaoui, Anouar Ben Mabrouk, and Carlo Cattani
CRC Press is an imprint of Taylor & Francis Group, LLC
The right of Sabrine Arfaoui, Anouar Ben Mabrouk, and Carlo Cattani to be identified as authors of this work has been
asserted by them in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers
have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright
holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowl-
edged please write and let us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or
utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including pho-
tocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission
from the publishers.
For permission to photocopy or use material electronically from this work, access www.copyright.com or contact the
Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. For works that are
not available on CCC please contact mpkbookspermissions@tandf.co.uk
Trademark notice: Product or corporate names may be trademarks or registered trademarks and are used only for
identification and explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
Names: Arfaoui, Sabrine, author. | Ben Mabrouk, Anouar, author. | Cattani,

Carlo, 1954- author.
Title: Wavelet analysis : basic concepts and applications / Sabrine
Arfaoui, University of Monastir, Anouar Ben Mabrouk, University of
Kairouan, Carlo Cattani, University of Tuscia.
Description: Boca Raton : Chapman & Hall/CRC Press, 2021. | Includes
bibliographical references and index.
Identifiers: LCCN 2020050757 (print) | LCCN 2020050758 (ebook) | ISBN
9780367562182 (hardback) | ISBN 9781003096924 (ebook)
Subjects: LCSH: Wavelets (Mathematics)
Classification: LCC QA403.3 .A74 2021 (print) | LCC QA403.3 (ebook) | DDC
515/.2433--dc23
LC record available at https://lccn.loc.gov/2020050757
LC ebook record available at https://lccn.loc.gov/2020050758
ISBN: 978-0-367-56218-2 (hbk)

ISBN: 978-1-003-09692-4 (ebk)
Typeset in Latin Modern font

by KnowledgeWorks Global Ltd.
Contents
List of Figures ix
Preface xi
Chapter 1 Introduction 1
Chapter 2 Wavelets on Euclidean Spaces 5
2.1 INTRODUCTION 5
2.2 WAVELETS ON R 6
2.2.1 Continuous wavelet transform 7
2.2.2 Discrete wavelet transform 10
2.3 MULTI-RESOLUTION ANALYSIS 11
2.4 WAVELET ALGORITHMS 13
2.5 WAVELET BASIS 16
2.6 MULTIDIMENSIONAL REAL WAVELETS 21
2.7 EXAMPLES OF WAVELET FUNCTIONS AND MRA 22
2.7.1 Haar wavelet 22
2.7.2 Faber–Schauder wavelet 24
2.7.3 Daubechies wavelets 25
2.7.4 Symlet wavelets 27
2.7.5 Spline wavelets 27
2.7.6 Anisotropic wavelets 29
2.7.7 Cauchy wavelets 30
2.8 EXERCISES 31
Chapter 3 Wavelets extended 35
3.1 AFFINE GROUP WAVELETS 35

3.2 MULTIRESOLUTION ANALYSIS ON THE INTERVAL 37
v
vi Contents
3.2.1 Monasse–Perrier construction 37

3.2.2 Bertoluzza–Falletta construction 37
3.2.3 Daubechies wavelets versus Bertoluzza–Faletta 39
3.3 WAVELETS ON THE SPHERE 40
3.3.1 Introduction 40
3.3.2 Existence of scaling functions 41
3.3.3 Multiresolution analysis on the sphere 43
3.3.4 Existence of the mother wavelet 44
3.4 EXERCISES 47
Chapter 4 Clifford wavelets 51
4.1 INTRODUCTION 51
4.2 DIFFERENT CONSTRUCTIONS OF CLIFFORD ALGEBRAS 52
4.2.1 Clifford original construction 53
4.2.2 Quadratic form-based construction 53
4.2.3 A standard construction 54
4.3 GRADUATION IN CLIFFORD ALGEBRAS 56
4.4 SOME USEFUL OPERATIONS ON CLIFFORD ALGEBRAS 57
4.4.1 Products in Clifford algebras 57
4.4.2 Involutions on a Clifford algebra 58
4.5 CLIFFORD FUNCTIONAL ANALYSIS 60
4.6 EXISTENCE OF MONOGENIC EXTENSIONS 67
4.7 CLIFFORD-FOURIER TRANSFORM 70
4.8 CLIFFORD WAVELET ANALYSIS 76
4.8.1 Spin-group based Clifford wavelets 76
4.8.2 Monogenic polynomial-based Clifford wavelets 82
4.9 SOME EXPERIMENTATIONS 92
4.10 EXERCISES 96
Chapter 5 Quantum wavelets 99
5.1 INTRODUCTION 99
5.2 BESSEL FUNCTIONS 99
5.3 BESSEL WAVELETS 105
5.4 FRACTIONAL BESSEL WAVELETS 107
5.5 QUANTUM THEORY TOOLKIT 119
Contents vii
5.6 SOME QUANTUM SPECIAL FUNCTIONS 123

5.7 QUANTUM WAVELETS 127
5.8 EXERCISES 134
Chapter 6 Wavelets in statistics 137
6.1 INTRODUCTION 137

6.2 WAVELET ANALYSIS OF TIME SERIES 138
6.2.1 Wavelet time series decomposition 138
6.2.2 The wavelet decomposition sample 140
6.3 WAVELET VARIANCE AND COVARIANCE 141
6.4 WAVELET DECIMATED AND STATIONARY TRANSFORMS 144
6.4.1 Decimated wavelet transform 144
6.4.2 Wavelet stationary transform 145
6.5 WAVELET DENSITY ESTIMATION 145
6.5.1 Orthogonal series for density estimation 145
6.5.2 δ-series estimators of density 147
6.5.3 Linear estimators 148
6.5.4 Donoho estimator 150
6.5.5 Hall-Patil estimator 150
6.5.6 Positive density estimators 151
6.6 WAVELET THRESHOLDING 152
6.6.1 Linear case 152
6.6.2 General case 154
6.6.3 Local thresholding 155
6.6.4 Global thresholding 155
6.6.5 Block thresholding 156
6.6.6 Sequences thresholding 156
6.7 APPLICATION TO WAVELET DENSITY ESTIMATIONS 157
6.7.1 Gaussian law estimation 158
6.7.2 Claw density wavelet estimators 159
6.8 EXERCISES 160
Chapter 7 Wavelets for partial differential equations 163

7.2 WAVELET COLLOCATION METHOD 165
viii Contents
7.3 WAVELET GALERKIN APPROACH 166

7.4 REDUCTION OF THE CONNECTION COEFFICIENTS NUMBER 171
7.5 TWO MAIN APPLICATIONS IN SOLVING PDEs 174
7.5.1 The Dirichlet Problem 174
7.5.2 The Neumann Problem 176
7.6 APPENDIX 179
7.7 EXERCISES 180
Chapter 8 Wavelets for fractal and multifractal functions 183

8.2 HAUSDORFF MEASURE AND DIMENSION 184
8.3 WAVELETS FOR THE REGULARITY OF FUNCTIONS 186
8.4 THE MULTIFRACTAL FORMALISM 189
8.4.1 Frisch and Parisi multifractal formalism conjecture 189
8.4.2 Arneodo et al wavelet-based multifractal formalism 190
8.5 SELF-SIMILAR-TYPE FUNCTIONS 192
8.6 APPLICATION TO FINANCIAL INDEX MODELING 201
8.7 APPENDIX 205
8.8 EXERCISES 205
Bibliography 209
Index 237
List of Figures
2.1 Decomposition algorithm. 15

2.2 Inverse algorithm. 16
2.3 Haar scaling and wavelet functions. 24
2.4 Schauder scaling and wavelet functions. 26
2.5 Some Daubechies scaling functions and associated wavelets. 27
2.6 Some symlet scaling functions and associated wavelets. 28
4.1 Wedge product of vectors. 58

4.2 Some 2D Clifford wavelets. 93
4.3 ψ1,2 (x)-Clifford wavelet 1-level decomposition of Mohamed Amine’s
photo. 94
4.4 The 3D ψ3,3 (x) Clifford wavelet brain processing at the level J = 2. 95
5.1 Graphs of Bessel functions Jv of the first kind for v = 0, 1, 2, 3. 102

5.2 Graphs of Bessel functions Yv of the second kind for v = 0, 1, 2, 3. 103
6.1 The normal reduced and centered density N (0, 1). 158
6.2 The gaussian density and its wavelet estimator. 159
6.3 Claw density. 159
6.4 Claw density wavelet estimator at the level J = 4. 160
8.1 Wavelet decomposition of the SP500 signal at the level J = 4. 203

8.2 SP500 spectrum of singularities. 204
8.3 Original SP500 and self-similar-type model. 204
ix
Preface
Nowadays, wavelets are applied almost everywhere in science. Both pure fields, such
as mathematics and theoretical physics, and applied ones, such as signal/image pro-
cessing, finance and engineering, apply wavelets. Although the references and/or the
documentation about wavelets and their applications are wide, it seems that with the
advancement of technology and the appearance of many phenomena in nature and in
life there still exist some places for more efforts and developments to understand the
new problems, as the existing wavelet methods do not provide good understanding
of them. The new COVID-19 pandemic may be one of the challenges that should be
understood.
On the other hand, especially for young researchers, existing references such as
books in wavelet theory are somehow very restricted. The majority are written for
specific communities. This is, in fact, not surprising and may be due to the necessity
of developing such references to overcome the concerned problems in that time.
Next, with the inclusion of wavelet theory in academic studies such as in mas-
ter’s and PhD programs, the scientific and academic communities have had a great
need to develop references in other forms. Students and generally researchers need
sometimes self-containing references responding to their need, to avoid losing time
in redeveloping existing results, which is a necessary step for both the generalization
and the experiments.
The present volume is composed of eight chapters. In the first introductory chap-
ter, a literal introduction is developed discussing generally the topic. Chapter 2 is
concerned with the presentation of the original developments of wavelet theory on
the real Euclidean space. This is also a preliminary chapter that will be of great help
for young researchers. Chapter 3 is more specialized and constitutes a continuation of
the previous one, in which some extending cases of wavelet theory and applications
have been provided. Chapter 4 is a very specialized part that is developed for the
first time to our knowledge. It is concerned with the presentation of wavelet theory in
a general functional framework based on Clifford algebras. This is very important as
these algebras contain all the Euclidean structures and gather them in one structure
to facilitate calculus. Readers will notice clearly that Clifford wavelet theory induces
naturally the Euclidean ones such as real and complex numbers, circles and spheres.
Chapter 5 is a continuation of the development of the theory in specialized fields
such as quantum theory. Next, in Chapter 6, statistical application of wavelets has
been reviewed. Topics such as density estimation, thresholding concepts, variance and
covariance have been detailed. Chapter 7 is devoted to wavelets applied in solving
partial differential equations. Recall that this field needs many assumptions on the
functional bases applied, especially the explicit form of the basis elements and their
xi
xii Preface
regularities. The last chapter is devoted to the link and/or the use of wavelet theory
in characterizing fractal and multifractal functions and their application. Each chap-
ter contains a series of exercises and experimentations to help understand the theory
and also to show the utility of wavelets.
The present book stems, in fact, from lectures and papers on the topics developed,
which have been gathered, re-developed, improved and sometimes completed with
necessary missing developments. However, naturally it is not exhaustive and should
be always criticized, sometimes corrected and improved by readers. So, we accept and
wait for any comments and suggestions.
We also want to stress the fact that we have provided in some chapters, especially
those on preliminary concepts that may be useful to young researchers, some exercises
and applications that are simple to handle with the aim to help the readers understand
the theory. We apologize if there are simpler applications and details that may be
more helpful to the readers but that have been left out from inclusion in this book.
This, in fact, needs more time and may induce delays in the publication of the book.
We hope that with the present form the readers become acquainted with the topics
presented.
The aim of this book is to provide a basic and self-contained introduction to
the ideas underpinning wavelet theory and its diversified applications. Readers of
our proposed book would include master’s degree students, PhD students and se-
nior researchers. It may also serve scientists and research workers from industrial
settings, where modeling real-world phenomena and data needs wavelets such as fi-
nance, medicine, engineering, transport, images and signals. Henceforth, the book
will interest practitioners and theorists alike. For theorists, rigorous mathematical
developments will be presented with necessary prerequisites that make the book self-
containing. For the practitioner, often interested in model building and analysis, we
provide the cornerstone ideas.
As with any scientific production and reference, the present volume could not have
been realized without the help of many persons. We thus owe thanks to many persons
who have helped us in any direction such as encouragements, scientific discussions
and documentation. We thank the Taylor & Francis Publishing Group for giving us
the opportunity to write and publish the present work. We also would like to express
our gratitude to our professors, teachers, colleagues, and universities. Without their
help and efforts, no such work might be realized. We would also like to thank all the
members of the publishing house, especially the editorial staff for the present volume,
Callum Fraser and Mansi Kabra, for their hospitality, cooperation, collaboration and
for the time they have spent on our project.
CHAPTER 1
Introduction
Wavelets were discovered in the eighteenth century, essentially in pertroleum extrac-

tion. They have induced a new type of analysis extending the Fourier one. Recall that
Fourier analysis has been for long a time the essential mathematical tool, especially in
harmonic analysis and related applications such as physics, engineering, signal/image
processing and PDEs. Next, wavelets were introduced as new extending mathemat-
ical tools to generalize the Fourier one and to overcome, in some ways, the disad-
vantages of Fourier analysis. For a large community, especially non-mathematicians
and non-physicists, a wavelet may be defined in the most simple sense as a wave
function that decays rapidly and has a zero mean.
Compared to the Fourier theory, wavelets are mathematical functions permitting
themselves to cut up data into different components relative to the frequency spec-
trum and next focus on these components somehow independently, extracting their
characteristics and lifting to the original data. One main advantage of wavelets is the
fact that they are more able than Fourier modes to analyze discontinuities and/or
singularities efficiently (see [20], [177], [218], [248], [279], [295], [305]).
Wavelets have been also developed independently in the fields of mathematics,
quantum physics, electrical engineering and seismic geology. Next, interchanges be-
tween these fields have yielded more understanding of the theory and more and more
bases as well as applications such as image compression, turbulence, human vision,
radar and earthquake.
Nowadays, wavelets have become reputable and successful tools in quasi all do-
mains. The particularity in a wavelet basis is that all its elements are deduced from
one source function known as the mother wavelet. Next, such a mother gives rise
to all the elements necessary to analyze objects by simple actions of translation,
dilatation and rotation. The last parameter was introduced by Antoine and his col-
laborators ([11], [13]) to obtain some directional selectivity of the wavelet transform
in higher dimensions. Indeed, unlike Fourier analysis, there are different ways to de-
fine multidimensional wavelets. Directional wavelets based on the rotation parameter
just evoked. Another class is based on tensor products of one-dimensional wavelets.
Also, some wavelets are related to manifolds, essentially spheres, where the idea is
based on the geometric structure of the surface where the data lies. This gives rise to
1
2 Wavelet Analysis: Basic Concepts and Applications
the so-called isotropic and anisotropic wavelets (see [11], [12], [20], [205], [218], [219],
[220], [229], [230], [248], [279], [295], [305], [357], [358], [363]).
Wavelet theory provides for functional spaces and time series good bases, allow-
ing their decomposition into spaces associated with different horizons known as the
levels of decomposition. A wavelet basis is a family of functions obtained from one
function known as the mother wavelet, by translations and dilations. Due to the
power of their theory, wavelets have many applications in different domains such as
mathematics, physics, electrical engineering and seismic geology. This tool permits
the representation of L2 -functions in a basis well localized in time and in frequency.
Wavelets are also associated with many special functions such as orthogonal poly-
nomials and hypergeometric series. The most well known may be the Bessel functions
that have been developed in both classic theory of Bessel functional analysis and the
modified versions in fractional and quantum calculus. As its name indicates, Bessel
wavelets are related to Bessel special function. Historically, special functions differ
from elementary ones such as powers, roots, trigonometric and their inverses, mainly
with the limitations that these latter classes are known for. Many fundamental prob-
lems such as orbital motion, simultaneous oscillatory chains and spherical body grav-
itational potential were not best described using elementary functions. This makes
it necessary to extend elementary functions’ classes to more general ones that may
describe well unresolved problems.
Wavelets are also developed and applied in financial time series such as market
indices and exchange rates. In [42], for example, a study of the largest transaction
financial market was carried out. The exchange market gave some high-frequency
data. Compared to other markets, such data can be available at long periods and
with high frequency. The data were detected for very small periods, which means
that the market is also liquid. Until 1990, economists were interested in intra-daily
data because of which the detection of some behaviors did not appear in the daily
analysis of data such as homogeneity.
A well-known hypothesis in finance is the homogeneity of markets where all in-
vestigators have almost the same behavior. The idea of nonhomogeneous markets is
more recent, and it suggests that investigators have different perceptions and differ-
ent laws. For the exchange market, for example, investigators can differ in profiles,
geographic localizations and also in institutional constraints. Another natural sug-
gestion can be done about traders. Naturally, traders investigating at short time in-
tervals allow some high-frequency behaviors in the change market. Long-time traders
are interested in the general tendency and the volatility of the market along a mi-
croscopic greed. Short-time traders, however, are interested in fractional perceptions
and so in macroscopic greed. This leads to the wavelet analysis of financial time
series.
Recently, other models have been introduced in modeling financial time series
by means of fractals, which are in turn strongly related to wavelets. For example,
in Olsen & Associates, operating the largest financial database, has noticed that the
tick frequency has strongly increased in one decade, causing problems in studying the
time series extracted from such a database. Such problems can be due to transmission
delays, input errors and machine damages. So, some filtering procedure has to be done
Introduction 3
before using the data. The first point that one must take into account in filtering time
series is their scaling behavior. Scaling laws were empirically observed by Olsen et al
until 1990 (see [306]). A time series X(t) has a scaling law if its so-called partition
function has the form
N/t
t X
Sq (t) = |X(jt)|q ∼ tτq +1
N j=1
where N stands for the size of the series for some appropriate function τ . This estima-
tion is well understood when merging wavelet tools into fractal models. (See Chapter
8.) In such models, Fourier analysis could not produce good results in estimating such
behavior. Indeed, Fourier transform of time series is generally limited because a single
analysis window cannot detect features in the signals that are either much longer or
shorter than the window size. Moving-window Fourier transform (MWFT) slides a
fixed-size analysis window along the time axis and is able to detect non-stationarities.
The fixed-size window algorithm of MWFT limits the detection of cycles at wave-
lengths that are longer than the analysis windows, and non-stationarities in short
wavelengths (i.e., high frequencies) are smoothed. Use of the wavelet transform solves
this problem, because it uses narrow windows at high frequencies and wide windows
at low frequencies.
This book is devoted to developing the basic concepts of wavelet analysis neces-
sary for young researchers doing their Master’s level in science and researchers doing
doctoral studies in pure mathematical/physical sciences, as well as applied and inter-
acted ones by providing the basic tools required, with simple and rigorous methods.
It also aims to serve researchers at advanced levels by providing them the necessary
tools that will allow them to understand and adapt wavelet theory to their needs
such as supervision and development of research projects.
The book provides some highly flexible methods and ideas that can be manip-
ulated easily by undergraduate students, and thus may be of interest for Bachelors
in science by providing them a clear idea on what wavelets are, and thus permitting
them to decide in their scientific future.
Organization of the book

Chapter 2 presents the notion of wavelets as analyzing functions and as mathematical
tools for analyzing square integrable functions known in signal theory as finite vari-
ance and/or finite energy signals. The analysis passes through two essential types of
transforms: the continuous wavelet transforms and the discrete wavelet transforms.
Such transforms are applied to represent the analyzed signals by means of wavelet
series in the time–frequency domain and thus use the modes generating such series
to localize singularities. Furthermore, we review multi-resolution analysis as a basic
construction tool related to wavelets. It allows to split the whole space of analyzed sig-
nals into sub-spaces known as approximation spaces and detail ones. Multi-resolution
analysis is based on nested sub-spaces that are related to each other by means of spe-
cific algorithms such as the decomposition low-pass filter-based algorithm and the
inverse high-pass one. These algorithms are efficient for understanding the behavior
of a series and eventual predictions.
Chapter 3 is a mixture of different concepts. We aim precisely to develop it as a

continuation of the previous one in order to show some extensions of wavelet theory
and also some other point of view to its theoretical introduction, essentially affine
group method, and also to show to the readers that although manifolds are parts of
Euclidean spaces for many cases, the concept of introducing wavelets on them may
differ. We considered the special case of spheres.
Chapter 4 is a new developed academic reference devoted to Clifford wavelet anal-
ysis. It offers a general context of Euclidean wavelet analysis by a higher-dimensional
analogue. The notion of monogenic function theory is reviewed: monogenic poly-
nomials and their application to yield Clifford wavelets. Mathematical formulations
of harmonic analysis such as Fourier–Plancherel and Parseval are established in the
new context. Applications in image processing are also developed. Readers will notice
clearly that Clifford wavelet theory induces naturally the Euclidean ones such as real
and complex circles and spheres.
Chapter 5 is a continuation of the development of the theory in specialized fields
such as quantum theory. We precisely present in detail quantum and fractional Bessel
functions and associated wavelet theory. Plancherel/Parseval as well as reconstruc-
tion formula has been investigated. Bessel wavelets are applied in various domains,
especially partial differential equations, wave motion, diffusion, etc.
Next, in Chapter 6, statistical application of wavelets is reviewed. Density esti-
mation, thresholding concepts, and variance and covariance topics are discussed in
detail. Recall that statistical and time series constitute a very delicate area of study
due to specific characteristics. Most of the time-varying series are nonlinear, in partic-
ular, the financial and economic series, which present an intellectual challenge. Their
behavior seems to change dramatically, and uncertainty is always present. To un-
derstand and to discover hidden characteristics and behavior of these series, wavelet
theory has been proved to be useful and necessary compared to existing previous
tools in statistics.
Chapter 7 is devoted to wavelets applied in solving partial differential equations.
Recall that this field needs many assumptions on the functional bases applied, espe-
cially the explicit form of the basis elements and their regularities. We propose to
show the contribution of wavelet theory in PDEs solving.
The idea is generally not complicated, and it consists essentially of developing the
unknown solution of the PDE into its eventual wavelet series decomposition (and its
derivatives included in the PDE) and then using the concept of wavelet basis to obtain
algebraic and/or matrices/vector equations on the wavelet coefficients that should be
resolved. The crucial point in this theory is the so-called connection coefficients of
wavelets. These have been investigated by many authors. In the present work, we
develop a somehow new procedure to compute them. Some applications are also
provided at the end of the chapter.
The last chapter is devoted to the link and/or the use of wavelet theory in
characterizing fractal and multifractal functions and their applications. Concepts
such as Hölder regularity, spectrum of singularity, multifractal formalism for func-
tion and self-similar-type functions based on wavelets are discussed with necessary
developments.
CHAPTER 2
Wavelets on Euclidean
Spaces
2.1 INTRODUCTION
Wavelets were discovered in the eighteenth century in petroleum exploration, and
since their discovery, they have proven to be powerful tools in many fields from
pure mathematics to physics to applied ones such as images, signals, medicine, fi-
nance and statistics. The study of their constructions and their properties, especially
in functional/signals decompositions on functional wavelet bases, has indeed grown
considerably.
In pure mathematics, wavelets constitute a refinement of Fourier analysis as they
compensate and/or resolve some anomalies in Fourier series. The first wavelet basis
has been, in fact, used before the pure mathematical discovery of wavelet bases since
the introduction of Haar system, which by the next has been proved indeed to be
a possible wavelet basis reminiscent of the regularity. Such a system dates back to
the beginning of the 20th century and was precisely discovered in the year 1909. It
was introduced in order to construct a functional basis permitting the representation
of all continuous functions by means of a uniformly convergent series. Recall that in
Fourier series, there are, as usual, many kinds of convergence that may be investigated
such as the point-wise convergence subject of the well-known Dirichlet theorem, the
uniform convergence which needs more assumptions on the series and the function,
the convergence in norm and Carleson’s or almost everywhere convergence. Each kind
of convergence requires special assumptions on the function. Although a Fourier series
converges in sense of Dirichlet, it does not imply that the graph of the partial Fourier
series converges to that of the function. This phenomenon is known as the Gibbs
phenomenon and is related to the presence of oscillations in the Fourier series near
the discontinuity points of the function. This means that the uniform convergence
is not sufficient. One of the challenging concepts in wavelet analysis is its ability to
describe well the behavior of the analyzed function near its singularities and join or
more precisely extend the notion of Littlewood-Palay decomposition.
To overcome some drawbacks of Fourier analysis, mathematicians have intro-
duced a bit of modification called the windowed Fourier transform by computing the
5
original Fourier transform of the analyzed signals on a special localized extra function
called the window. However, some situations remain non resolved especially with the
emergence of irregular signals or high-frequency variations. The major problem in the
windowed Fourier extension is due to the use of fixed window, which may not be well
adapted to other problems such as high fluctuations of non stationary signals. This
led researchers to think about a stronger tool taking into account nonlinear algo-
rithms, nonstationary signals, nonperiodical, volatile and/or fluctuated ones. It holds
that wavelets, since their discovery, have permitted to overcome these obstacles.
These powers are related simultaneously to many properties of wavelets. Indeed,
wavelet decomposition of functions joins Littlewood-Paley decomposition in many
cases. Wavelets provide simultaneous local analyses related to time–frequency. They
can be adapted to study-specific operators, especially differential and stochastic ones.
From the numerical and/or applied point of view, wavelets provide fast and accurate
algorithms, multi-resolution analyses as well as recursive schemes. These are very
important especially in big data analysis, image processing and also in numerical
resolution of partial differential equations.
In this chapter, we propose to review the basic concepts of wavelets as well as
their basic properties.
2.2 WAVELETS ON R
The first wavelet bases and thus analyses have been constructed on the real line R and
have been next extended to the general cases of the real/complex Euclidean spaces
Rm -Cm using different methods such as the natural tensor product.
Mathematically speaking, a wavelet or an analyzing wavelet on the Euclidean
space Rm may be defined in a large way as a function with specific properties that
may or may not be required necessarily as simultaneous assumptions. More precisely,
wavelet analysis is based primarily on the following points:
• An effective representation for standard functions,
• Robustness to the specification models,
• A reduction in the computation time,
• Simplicity of the analysis,
• An easy generalization and efficient, depending on the dimension,
• A location in time and frequency.
A wavelet is a function ψ ∈ L2 (R) that satisfies the following conditions:
• Admissibility,
dω
Z
|ψ̂(ω)|2 = Cψ < ∞. (2.1)
R+ |ω|
• Zero mean, Z +∞
ψ(0)
b = ψ(u)du = 0. (2.2)
−∞
Wavelets on Euclidean Spaces 7
• Localization in time/frequency domains

Z +∞
ψ(0)
b = |ψ(u)|2 du = 1. (2.3)
−∞
• Enough vanishing moments,

Z
p = 0, ..., m − 1, ψ(t)tp dt = 0. (2.4)
R
To analyze a signal by wavelets, one passes as in Fourier analysis by the wavelet

transform of the signal. A wavelet transform (WT) is a re representation of a time–
frequency signal. It replaces the Fourier sine by a wavelet. Generally, there are two
types of processing: the continuous wavelet transform (CWT) and the discrete wavelet
transform.
2.2.1 Continuous wavelet transform

The CWT is based firstly on the introduction of a translation parameter u ∈ R and
another parameter s > 0 known as the scale to the analyzing wavelet ψ, which plays
the role of Fourier sine and cosine and will be subsequently called mother wavelet.
The translation parameter determines the position or the time around which we want
to assess the behavior of the signal, while the scale factor is used to assess the signal
behavior around the position. That is, it allows us to estimate the frequency of the
signal at that point. Let
1 x−u

ψs,u (x) = √ ψ . (2.5)
s s
The CWT at the position u and the scale s is defined by
Z ∞
du,s (f ) = ψu,s (t)f (t)dt, ∀ u, s. (2.6)
−∞
By varying the parameters s and u, we may cover completely all the time–frequency
plane. This gives a full and redundant representation of the whole signal to be an-
alyzed (see [295]). This transform is called continuous because of the nature of the
parameters s and u that may operate at all levels and positions. The original signal
S can be reproduced knowing its CWT by the following relationship:
1 x−u dsdu
Z Z
S(t) = du,s (S)ψ . (2.7)
Cψ R s s2
It remains to notice that CWT is suitable for continuous time signals and those
representing varying singularities.
The function ψ ∈ L2 (R) satisfies some conditions such as the admissibility con-
dition and somehow describes Fourier-Plancherel identity and says that
dω
Z
|ψ̂(ω)|2 = Cψ < ∞. (2.8)
R+ |ω|
The function ψ has to also satisfy a number of vanishing moments, which is related
in wavelet theory to its regularity order. It states that
Z
p = 0, ..., m − 1, ψ(t)tp dt = 0. (2.9)
R
Sometimes, we say that ψ is C m on R. The time localization chart is a normalization

form that is resumed in the identity
Z +∞
|ψ(u)|2 du = 1. (2.10)
−∞
To analyze a signal by wavelets, one passes via the so-called wavelet transforms. A
wavelet transform is a representation of the signal by means of an integral form similar
to Fourier one in which the Fourier sine and/or cosine is replaced by the analyzing
wavelet ψ. In Fourier transform, the complex exponential source function yields the
copies eis. indexed by s ∈ R, which somehow represent frequencies. This transform
is continuous in the sense that it is indexed on the whole line of indices s ∈ R.
In wavelet theory, the situation is more unified. A CWT is also well known.
Firstly, a frequency, scale or a dilation or compression parameter s > 0 and a second
one related to time or position u ∈ R have to be fixed. The source function ψ
known as the analyzing wavelet is next transformed to yield some copies (replacing
the eis. ):
1 x−u

ψs,u (x) = √ ψ . (2.11)
s s
The CWT of a real-valued function f defined on the real line at the position u and
the scale s is defined by
Z ∞
ds,u (f ) = f (t)ψs,u (t)dt, ∀ u, s. (2.12)
−∞
By varying the parameters s and u, we cover completely all the time–frequency plane.
This gives a full and redundant representation of the whole signal to be analyzed (see
[295]). This transform is called continuous because of the nature of the parameters s
and u that may operate at all levels and positions.
So, wavelets operate according two parameters: the parameter u, which permits
translation of the graph of the source mother wavelet ψ, and the parameter s, which
permits compression or dilation of the graph of ψ. Computing or evaluating the
coefficients du,s means analyzing the function f with wavelets.
Theorem 2.1 The wavelet transform ds,u (f ) possesses some properties such as
1. The linearity, in the sense that
ds,u (αf + βg) = αds,u (f ) + βds,u (g), ∀f, g.

2. The translation-invariance, in the sense that
ds,u (τt f ) = ds,u−t (f ), ∀f ; and ∀u, s, t,
and where
(τt f )(x) = f (x − t).
3. The dilation-invariance, in the sense that

1
ds,u (fa ) = √ das,au (f ), ∀f ; and ∀u, s, a,
a
and where for a > 0,

(fa )(x) = f (ax).
The proof of these properties is easy and the readers may refer to [163] for a review.
In wavelet theory, as in Fourier analysis theory, the original function f can be
reproduced via its CWT by an L2 -identity.
Theorem 2.2 For all f ∈ L2 (R), we have the L2 -equality
1 x − u dsdu
Z Z
f (x) = ds,u (f )ψ( ) 2 .
Cψ s s
The proof of this result is based on the following lemma:
Lemma 2.3 Under the hypothesis of Theorem (2.2), we have
dsdu
Z Z Z
ds,u (f )ds,u (g) = Cψ f (x)g(x)dx, ∀ f, g ∈ L2 (R).
s
Proof. We have
1 x−u 1
Z
{ fˆ(y)ψ̂(sy)e−iuy .

ds,u (f ) = f ∗ ψs (u) = √ f (x)ψ( )dx =
s s 2π
Consequently
1
Z Z
ds,u (f )ds,u (g)du = fb(y)ĝ(y)|ψ̂(sy)|2 dy.
u 2π y
By application of Fubini’s rule, we get

dsdu 1 dsdy
Z Z Z Z
ds,u (f )ds,u (g) = fˆ(y)ĝ(y)|ψ̂(sy)|2
s>0 u s 2π s>0 y s
1
Z
= Cψ fˆ(y)ĝ(y)dy
2π y
Z
= Cψ f (y)g(y)dy.
y
Proof of Theorem 2.2. By applying the Riesz rule, we get

1 x − b dadb
Z Z
||F (x) − Ca,b (F )ψ( ) 2 ||L2
Cψ 1/A≤a≤A |b|≤B a a
!
1 x − b dadb
Z Z Z
= sup F (x) − Ca,b (F )ψ( ) 2 G(x)dx.
||G||=1 Cψ 1/A≤a≤A |b|≤B a a
Next, using Fubini’s rule, we observe that the last line is equal to
1 dadb
Z Z Z
sup F (x)G(x)dx − Ca,b (F )Ca,b (G)
||G||=1 Cψ 1/A≤a≤A |b|≤B a
1 dadb
Z
= sup Ca,b (F )Ca,b (G)
||G||=1 Cψ (a,b)∈[1/A,A]×[−B,B]
/ a
which by Cauchy–Schwarz inequality is bounded by

"Z #1/2
1 2 dadb
|Ca,b (F )|
Cψ (a,b)∈[1/A,A]×[−B,B]
/ a
" 1/2
2 dadb
Z
sup |Ca,b (G)| .
||G||=1 (a,b)∈[1/A,A]×[−B,B]
/ a
Now, Lemma 2.3 shows that the last quantity goes to 0 as A, B tends to +∞.
2.2.2 Discrete wavelet transform

To analyze statistical series or discrete time signals and avoid redundancy problems
and integrals calculations appearing in the CWT, one makes use of the discrete
wavelet transform. It is to restrict to discrete calculations grids for scale parameters
and position instead of browsing the entire domain. The most used method is dyadic
grid based on taking s = 2−j and u = k2−j . In this case, the wavelet copy ψu,s will
be denoted by ψj,k and defined by
ψj,k (x) = 2−j/2 ψ(2j x − k), j, k ∈ Z.
The discrete wavelet transform will be defined by
Z ∞
dj,k = ψj,k (t)S(t)dt. (2.13)
−∞
These are often called wavelet coefficients or detail coefficients of the signal S.
It holds that the set (ψj,k )j,k∈Z constitutes an orthonormal basis of L2 (R) and it
is called wavelet basis. A signal S with finite energy (in L2 (R)) is then decomposed
according to this basis into a series
∞ X
X
S(t) = dj,k ψj,k (t) (2.14)
j=0 k
called the wavelet series of S which replaces the reconstruction formula for the CWT.
2.3 MULTI-RESOLUTION ANALYSIS

Multi-resolution analysis (MRA) is a functional framework for representing a series
of approximations to different levels called resolutions. MRA is a family of vector
spaces (Vj )j∈Z of the whole space of finite variance (energy) signals L2 (R) nested in
the sense of a scaling law. Recall that
Z
L2 (R) = {S : R −→ R; |S(t)|2 dt < ∞}. (2.15)
R
For each j ∈ Z, Vj is called the approximation space at the scale or the level j. More
precisely, we have the following definition ([295]).
Definition 2.4 A multiresolution analysis is a sequence of closed subsets (Vj )j∈Z of

L2 (R) that satisfies the following points.
a) ∀j ∈ Z; V0 ⊂ V1 ⊂ .... ⊂ Vj ⊂ Vj+1 .
b) ∀j ∈ Z; f ∈ Vj ⇔ f (2.) ∈ Vj+1
c) There exists ϕ ∈ V0 such that {ϕ0,k = ϕ(. − k); k ∈ Z} is a Riesz basis of V0 .

\
d) Vj = {0}.
j∈R
[
e) Vj = L2 (R).
j∈R
f) ∀j ∈ Z; f ∈ Vj ⇐⇒ f (x − k) ∈ Vj
The property (a) reflects that the approximation of a signal at the resolution’s level
j + 1 contains the necessary information to yield the approximation at the level j.
Assertion (b) called dilatation’s property permits passing from a level of resolution
to another. The space Vj+1 contains signals that are coarser than Vj . Assertion (c)
means that a scaling function exists and permits the decomposition of the signal at the
starting level 0. The property (d) means that at very low resolution level (2−j −→ 0
as j −→ +∞) we lose all the details of the signal. At a minimal resolution, we
lose all the information about the signal. Assertion (e) implies that the signal may
be approximated with elements in Vj . At a maximal resolution, we reconstruct all
the whole signal. Finally, the property (f) of translation means that the space Vj is
invariant under integer translation.
Definition 2.5 The source function ϕ is called the scaling function of the MRA or
also the father wavelet.
It holds that this function generates all the subspaces Vj ’s of the MRA by acting
as dilation/translation parameters. Indeed, the property
(d) combined with
(f) in
j/2 j
Definition 2.4 implies that, for all j, ∈ Z, the set ϕj,k (x) = 2 ϕ(2 x − k) is an
k
orthogonal basis of Vj . From assertion (a), it holds that we may complete Vj in Vj+1
in the sense of the direct sum. Let Wj be the orthogonal supplementary of Vj in Vj+1 ,
that is
Vj+1 = Vj ⊕⊥ Wj . (2.16)
We will see that Wj plays a primordial role in representing the details of the analyzed
signal. This is why it is called the detail space at the level j. Iterating the relation
(2.16), we obtain an orthogonal decomposition for all J ∈ Z,
J,⊥
VJ = V0 ⊕⊥
M
Wj . (2.17)
j=0
By exploiting (c), this leads to

M
L2 (R) = Wj . (2.18)
j∈Z
L2 (R) is decomposed into subspaces that are mutually orthogonal.

Definition 2.6 The space Wj , j ∈ Z, is called detail space at the scale or the level
j.
In wavelet theory, the following result is proved.
Theorem and Definition 2.7 There exists a function ψ ∈ W0 that satisfies
• (ψ(t − k))k is an orthogonal basis of W0 .

• ψj,k (t) = 2j/2 ψ(2j t − k) is an orthogonal basis of Wj .
k
The function ψ is called the wavelet function or the mother wavelet associated with
the scaling function ϕ of the MRA.
We will now explain the relationship between ϕ and ψ. It holds from (2.18) that any
S ∈ L2 (R) is decomposed into components according to the subspaces Wj ’s in the
sense that
+∞
X
S= S Wj ,
j=−∞
where the SWj for j ∈ Z designates the orthogonal

projection of S on W j
. Recall
next that Wj is generated by the orthogonal basis ψj,k (t) = 2j/2 ψ(2j t − k) for any
k
j ∈ Z. Consequently, X
SWj = < S, ψj,k > ψj,k .
k
Which leads to
+∞
X X
S= < S, ψj,k > ψj,k .
j=−∞ k
Observing now that the wavelet coefficients are already defined by (2.12) or (2.13) as
Z +∞
dj,k (S) = S(t)ψj,k (t)dt =< S, ψj,k >
−∞
we obtain the wavelet series of S defined already by (2.14). Returning next to (2.16)
or also (2.17), we may write that
−1 X
X +∞
XX
S= < S, ψj,k > ψj,k + < S, ψj,k > ψj,k . (2.19)
j=−∞ k j=0 k
It results from the nesting property of the Vj ’s and the fact that Wj ⊂ Vj+1
that the first part in (2.19) is an element of V0 which is generated by the basis
(ϕk (t) = ϕ(t − k))k . Consequently, we may also write that
−1 X
X X
< S, ψj,k > ψj,k = Ck ϕk .
j=−∞ k k
Hence, the wavelet series decomposition of S becomes

X ∞ X
X
S(x) = Ck ϕk (x) + dj,k ψj,k (x). (2.20)
k j=0 k
In fact, it may truncate at any level J and apply the ϕJ,k instead of ϕk to obtain
X ∞ X
X
S(x) = CJ,k ϕJ,k (x) + dj,k ψj,k (x) (2.21)
k j=J k
which is known as the wavelet series decomposition of S at the level J. In fact, the
coefficients that appears in this decomposition are evaluated via the relation
Z +∞
CJ,k = S(t)ϕJ,k (t)dt
−∞
and are called the scaling or approximation coefficients of S. The first component
in VJ reflects the global behavior or the tendency of the whole signal S and the
second component relative to the dj,k ’s represents the details of S and thus reflects
the dynamic behavior of the signal.
2.4 WAVELET ALGORITHMS

The strongest point in MRA and wavelet theory is that the scaling function and the
analyzing wavelet lead each one to the other. Indeed, recall that ϕ belongs to V0 ⊂ V1
and the latter is generated by the basis (ϕ1,k )k . Hence, ϕ is expressed by means of
(ϕ1,k )k . More precisely, we have the following result.
Theorem and Definition 2.8 • The scaling function satisfies the so-called
two-scale relation √ X
ϕ(x) = 2 hk ϕ(2x − k) (2.22)
k
where the coefficients hk are
Z
hk = ϕ(x)ϕ(2x − k)dx.
R
• The mother wavelet ψ is expressed as

√ X
ψ(x) = 2 gk ϕ(2x − k)
k
where the gk ’s are evaluated by
gk = (−1)k h1−k .
For more details, we refer to [177], [236], [295]. We will now prove that these relations
allow obtaining all the decomposition of a signal from each other through specific
algorithms. Indeed, consider a signal S and its approximation coefficients Cj,k and
details dj,k . For j ∈ Z, we have
Cj,k = hS, ϕj,k i.
It follows from the two-scale relation (2.22) that

X
ϕj,k = hl ϕj+1,l+2k .
l
Hence,
X
Cj,k = hl Cj+1,2k+l .
l
This means that the approximation at level j is obtained from the level j + 1 by the
intermediate of a filter. We have, in fact, the following definition.
Definition 2.9 The sequence H = (hk )k is called the discrete low-pass filter.
Analogously, we have for j, k fixed,

X
ψj,k (x) = gl ϕj+1,l+2k .
Hence, X
dj,k = hS, ψj,k i = gl Cj+1,l+2k .
l
This means that the approximation at level j is obtained from the level j + 1 by the
intermediate of a filter. We have here also the following definition.
Definition 2.10 The sequence G = (gk )k is called the discrete high-pass filter.
Figure 2.1 presents the decomposition algorithm due to Mallat [295]. It consists of a
cascade algorithm permitting to obtain all the levels of resolution.
We will now explain the inverse algorithm which permits to obtain the level j + 1
from the level j. Recall that the orthogonal projection of S on the approximation
space Vj is given by (Hardle et al (1997) [236])
X X
SVj = Cj,k ϕj,k + dj,k ψj,k
k k
Cj+1
H G
↓2 ↓2
Cj dj
H G
↓2 ↓2
Cj−1 dj−1
Figure 2.1 Decomposition algorithm.
and that
Cj+1,k = < SVj , ϕj+1,k >

X X
= Cj,k < ϕj,k , ϕj+1,k > + dj,k < ψj,k , ϕj+1,k >
k k
or else X X
Cj+1,n = hn−2k Cj,m + gn−2k dj,m . (2.23)
k k
This reconstruction is illustrated by means of Figure 2.2.

Finally, we recall some properties related to the filters H and G ([295]).
Proposition 2.11 The filters H and G satisfy

X
1. hn hn+2j = 0; ∀j 6= 0.
n
X
2. h2n = 1.
n
3. gn = (−1)n hl−n ; ∀ n.
X
4. hn gn+2j = 0, ∀ j ∈ N. (called mutual orthogonality).
n
Cj−1 dj−1
↑2 ↑2
H G
Cj dj
↑2 ↑2
H G
Cj+1
Figure 2.2 Inverse algorithm.
2.5 WAVELET BASIS

Providing functional spaces with bases is the most important task in functional anal-
ysis. It permits in some sense to reduce the problem of proving properties and/or
demanding characteristics from a function—usually unknown—to doing it and/or
demanding it from the elements of the basis. In finite dimensional spaces, this is
somehow equivalent. In general and especially for almost all functional spaces where
the dimension is usually infinite, we intend that the basis satisfy more properties to
permit the reduction.
In wavelet bases, for example, one of the important properties required is the
localization and regularity of the mother wavelet ψ, which is obviously inherited by
all the elements ψj,k .
Wavelet bases may also be related to the concept of multiresolution analysis where
an important property is also required and looks like the localization and consists
of the existence of Riesz basis to define a multiresolution analysis. Recall that in a
separable Hilbert space, a countable collection (ek )k∈Z is said to be a Riesz basis if
all its finite linear combinations are dense in the whole Hilbert space and in addition
if it satisfies an equivalence inequality of the form
X X X
K1 |λk |2 ≤ k λk ek k2 ≤ K1 |λk |2
k k k
for some positive constants K1 ≤ K2 .

The following result provides a characterization of the scaling function—father
wavelet—of a multiresolution analysis to be effectively a good candidate.
X
Proposition 2.12 Let ϕ ∈ L2 (R) and denote Γϕ (ξ) = b + 2kπ)|2 . Then, the
|ϕ(ξ
k∈Z
collection (ϕ(x − k))k∈Z is a Riesz basis for V0 if and only if
0 < inf Γϕ ≤ Γϕ (ξ) ≤ sup Γϕ < ∞, ∀ξ.
X (ϕ(x − k))k∈Z is a Riesz basis for V0 and let

Proof 2.1 Assume that the collection
(λk )k∈Z ∈ `2 (R). Denote Fϕ (x) = λk ϕ(x − k). Its Fourier transform is F
cϕ (ξ) =
X k∈Z
ikξ
λk e ϕ(ξ).
b We next have
k∈Z
Z X
cϕ k2 2
kF L (R) = | λk eikξ ϕ(ξ)|
b 2
dξ
R k
XZ 2π(l+1) X
= | λk eikξ ϕ(ξ)|
b 2
dξ
l∈Z 2πl k
X Z 2π X
= | λk eikξ ϕ(ξ
b + 2πl)|2 dξ
0
Zl∈Z2π X k
X
= | λk eikξ |2 ( b + 2πl)|2 )dξ.
|ϕ(ξ
0 k l∈Z
As a result,
1 c 2 1
Z 2π X
kFϕ k2L2 (R) = kFϕ kL2 (R) = | λk eikξ |2 Γϕ (ξ)dξ.
2π 2π 0 k
1 X
Now, observing that the map Φ : λ = (λk ) → Φ(λ) = √ λk eikξ is an iso-
2π
morphism from `2 (R) topL2 ([0, 2π]), to get the proposition, the map K defined on
L2 ([0, 2π]) by K(f ) = f Γϕ should also be an isomorphism. This is the case if and
1
only if Γϕ and are both bounded (in L∞ ([0, 2π])).
Γϕ
Sometimes we need more regularity on the wavelet basis especially in the case of par-
tial differential equations. Besides, we sometimes need compactly supported wavelets
and/or wavelets with fast decay which are more adaptable and suitable for boundary
and/or limit conditions due to partial differential equations.
It is sometimes necessary to assume that the scaling function ϕ and its derivatives
to an order r (usually the order of the MRA) are of fast decay. This permits to
avoid the assumption on the overlap function Γϕ of ϕ to be upper bounded, as this
characterization becomes already holding. Indeed, whenever ϕ is of fast decay, its
Fourier transform ϕb is therefore infinitely continuously differentiable and is also of
fast decay. It satisfies, in particular,
K
|ϕ(ξ)|
b ≤
1 + |ξ|
for some constant K. This leads to an upper-bounded overlap function Γϕ . As a con-

sequence of these facts, we may assume that Γϕ ≡ 1. This results in a correspondence
between `2 (R), which are r-order fast decay, and the elements of V0 , which are fast
decay, as well as their derivatives to the order r.
In some other applications, one may need more assumptions on the wavelet ba-
sis, especially in numerical analysis where the exponential decay behavior is more
suitable than just fast decay. It holds, in fact, that whenever the sequence (λk )k is
exponentially decaying in `2 (R), the associated function Fϕ is of exponential decay
in V0 . More generally, we may prove the following result.
X
Lemma 2.13 Let m(ω) = λk eikω . Then, m is analytic in some annulus |Imω| <
k
R (for some positive constant R) if and only if (|λk |)K has an exponential decay.
Proof 2.2 Assume that the trigonometric series m is analytic on the annulus
|Imω| < R. Observe next that m is also periodic with period 2π. As a consequence,
the function ω 7−→ m(ω + 2π) − m(ω) is vanishing on all the annulus. Now applying
the complex integral, we get
Z 2π Z −iR+2π
λk = m(ω)e−ikω dω = m(z)e−ikz dz
0 −iR
Observe next that

2π sup |m(z)|e−kImz ≤ Ce−kR .
z∈[−iR,iR+2π]
Consequently,
|λk | ≤ Ce−kR .
We will now show that in Definition 2.4 some properties and/or assumptions may
be deduced from others, which thus reduces the number of assumptions and gives
more flexibility in the introduction of multiresolution analyses. We have precisely the
following result.
Properties 2.14 Assume that assertions (a), (b) and (c) in Definition 2.4 hold and
denote Φ, the function defined by its Fourier transform
1
b + 2kπ)|2 )− 2 .
X
Φ(ω)
b = ϕ(ω)
b |ϕ(ω
k
It holds that Z +∞ X
ϕ(x)dx = 1 and ϕ(x − k) = 1.
−∞ k
Moreover, assertions (d), (e), and (f) in Definition 2.4 hold.
Proof 2.3 Observe from assertion (a) that

X j
ϕ(x) = hj,k 2 2 ϕ(2j x − k),
k
where the coefficients hj,k are the analogues of the filter coefficients hk in the two-scale
relation. More precisely,
Z +∞ 1
hj,k = ϕ(x)2 2 ϕ(2j x − k)dx.
−∞
Z +∞
Denote next Iϕ = ϕ(x)dx. We may write
+∞
k
Z +∞ k j
− 2j
hj,k − Iϕ 2 ϕ( j ) = hj,k − ϕ( )2 2 ϕ(2j x − k)dx
2 +∞ 2j
Z +∞ k j
≤ C |x − |2 2 |ϕ(2j x − k)|dx
+∞ Z 2j
+∞
−3 2j
≤ C2 |x||ϕ(x)|dx
+∞
−3 2j
≤ C2 .
As a result,
j k j
hj,k = Iϕ 2− 2 ϕ( j
) + O(2−3 2 ).
2
Consequently,
k j
)ϕ(2j x − k) + O(2−3 2 ).
X
ϕ(x) = Iϕ ϕ( j
k
2
We now claim that Iϕ = 1. Indeed, let x0 be such that ϕ(x0 ) 6= 0. Let δ, η > 0 be
such that |ϕ(x) − ϕ(x0 )| ≤ whenever |x − x0 | ≤ η. Therefore,
X X
ϕ(x0 ) − Iϕ ϕ(x0 )ϕ(2j x − k) ≤ Iϕ |ϕ(2j x − k)|
k | kj −x0 |≤η
2
k
X
+ |ϕ( )| + |ϕ(x0 )| |ϕ(2j x − k)|
k
2j
| −x0 |>η
2j
j
+C + (2−3 2 ).
η k
Observe next whenever |x − x0 | ≤ then we necessarily get | j − x0 | > η, which
2 2
k η
implies that | j − x| ≥ . Hence,
2 2
C C
|ϕ(2j x − k)| ≤ ≤ .
(1 + |2j x − k|)4 (1 + |2j x − k|)2 (2j η2 )2
As a consequence,
X X 1 −3 2j
ϕ(x0 ) − Iϕ ϕ(x0 )ϕ(2j x − k) ≤ C η 2 + C + O(2 )
k k
(1 + |2j x 2 j
− k|) (2 2 )
j
≤ C2 −2j −2
η + C + O(2−3 2 ).
It follows consequently that for j large enough

X
ϕ(x0 ) − Iϕ ϕ(x0 )ϕ(2j x − k) ≤ C.
k
This yields that M 6= 0, and whenever |y − 2j x0 | ≤ 2j η2 , we get
X 1
ϕ(y − k) − ≤ C.
k
Iϕ
X
Now observe that the function y 7−→ ϕ(y − k) is 1-periodic. Hence, for j large
k
enough, we get that such a function remains constant on the whole line R, which
means that
X 1
ϕ(y − k) = .
k
M
This yields
Z +∞ XZ k+1 Z 1 X 1
M= ϕ(y)dy = ϕ(y)dy = ( ϕ(y − k))dy = .
∞ k∈Z k 0 k∈Z
M
Hence, M 2 = 1 and \
thus may be chosen to be equal to 1.
We now prove that Vj = {0}. Let F be an element in the intersection. Hence, for
j
all j, we may write X
F = aj,k ϕj,k ,
k
where the aj,k are uniformly bounded. Moreover,
X
kF k2 = |aj,k |2 .
k
Therefore, for all j we get

X j
|F | ≤ C |ϕj,k | ≤ C2 2
k
which means that F = 0.

It remains now to show assertion (e) in Definition 2.4. To do this, we prove that any
piecewise constant function may be approximated by elements of the union of the
Vj ’s. So, let [a, b] be an interval of R and let for j ∈ Z the function
X
Fj = 2j/2 ϕj,k .
k
∈[a,b]
2j
We immediately observe that for all x we have

X 1 C
|Fj (x)| ≤ C ≤ j .
k
(1 + |2j x − k|2 ) 2 dist(x, [a, b])
∈[a,b]
2j
On the other hand,

X c
|1 − Fj (x)| = | ϕ(2j x − k)| ≤ .
k
2j dist(x, [a, b]c )
∈[a,b]
/
2j
Now observe that

Z Z
2
|χ[a,b] (x) − Fj (x)| dx = j |χ[a,b] (x) − Fj (x)|2 dx
dist(x,[a,b]c )≥2− 2
Z
+ −
j |χ[a,b] (x) − Fj (x)|2 dx
Zdist(x,[a,b])≥2 2
+ j |χ[a,b] (x) − Fj (x)|2 dx.

dist(x,∂[a,b])≤2− 2
Next, observing the estimations above we obtain

Z
j
|χ[a,b] (x) − Fj (x)|2 dx ≤ C2−2j 2j + C2−2j 2j + C2− 2 .
Consequently,
j
kχ[a,b] − Fj k22 ≤ C2− 2 .
Letting j → ∞ we observe Fj goes to χ[a,b] in L2 .
2.6 MULTIDIMENSIONAL REAL WAVELETS

Many ideas have been exploited to introduce multidimensional wavelet analysis. Some
are based on the adoption of multiresolution analysis on R to the multidimensional
case. The first constructed bases were separable ones. Their construction focus on an
analogy with the Haar one. Recall that in one-dimensional case, this basis is defined
by (
ψj,k (x) = 2j/2 ψ(2j x − k) ; j, k ∈ Z
ψ = ξ[0,1/2[ − ξ[1/2,1[
Generally, let (Vj1 ) be a multiresolution analysis of L2 (R) with a scaling function ϕ
and a wavelet ψ, and let the orthogonal projection on Vj be denoted by Pj . The main
idea in the multiresolution analysis is the ability to construct an orthonormal wavelet
basis {ψj,k ; j, k ∈ Z}, ψj,k (x) = 2j/2 ψ(2j x − k), such that
X
Pj+1 f = Pj f + < f, ψj,k > ψj,k ∀ f ∈, L2 (R).
k
The adoption of these one-dimensional MRA and wavelets will be the starting point
to construct the multidimensional case. Indeed, let (Vj1 ) be a multiresolution analysis
of L2 (R) with a scaling function ϕ, a wavelet ψ and an orthogonal projections Pj1 .
Consider then orthogonal projection Pjd in L2 (Rd ) defined as the tensor product of d
copies of Pj1
Pjd = Pj1 ⊗ Pj1 ⊗ ... ⊗ Pj1 .
Denote Vjd = Pjd (L2 (Rd )). We have Vjd = Vj1 ⊗ Vj1 ⊗ ... ⊗ Vj1 . The closure in L2 (Rd )
of Vjd has an orthonormal basis
(
ϕdj,k = 2jd/2 ϕd (2j x − k) ; j ∈ R , k ∈ Rd
ϕd = ϕ ⊗ ϕ ⊗ ... ⊗ ϕ
The detail spaces Wjd will be defined by
Wjd = ⊕ε6=(0,..,0) Vjε1 ⊗ Vjε2 ⊗ ... ⊗ Vjεd .
This yields an orthonormal basis of L2 (Rd ) associated with Pjd

(
ε
ψj,k (x) = 2jd/2 ψ ε1 (2j x1 − k1 )...ψ εd (2j xd − kd ) ; j, ki ∈ Z,
ψ 1 = ψ , ψ 0 = ϕ = ξ[0,1[ .
This last formula looks better than the one defined by tensor products and yields
simple separable wavelets. In contrast, nonseparable wavelets remain difficult to be
used and to construct. However, in analysis and in nature, one can speak about
propagations in privileged directions. One plans to study their behavior by means of
well-adapted wavelets. An important example of directional phenomena is supplied by
spirals, such as the domain between the two curves of equations (in polar coordinates)
r = θ−α and r = (θ + π)−α .
Another example that bears similarities with spirals is the set
[ 1 1

Cα = , .
n (2n + 1) (2n)α
α
If the aim is a pointwise analysis, without particular emphasis on directions, the

topic will be more economical. However, if the signal to be analyzed has a preferred
direction, then one needs a wavelet with good angular selectivity.
2.7 EXAMPLES OF WAVELET FUNCTIONS AND MRA

In this section, we propose to present some examples of wavelets and associated MRA.
The readers can be referred to [163], [248], [289], [295] for more details and examples
of original wavelet analysis on the real line and Euclidean spaces in general.
2.7.1 Haar wavelet

The example of Haar is the simplest example in the theory of wavelets and MRA. It
is based on the Haar scaling function explicitly given by
ϕ = χ[0,1[
and characterized by the possibility of explicit computations of the transforms and
coefficients. The basic approximation space is given by
( )
X X
V0 = f ∈ L2 (R) ; f = ak ϕk ; (ak )k ⊂ R such that a2k < ∞ .
k∈Z k∈Z
For k ∈ Z, we denote ϕk by the function

ϕk (t) = ϕ0,k (t) = χ[0,1[ (t − k) = χ[k,k+1[ (t).
Consequently, observing that the ϕk ’s have disjoints supports, any element S ∈ V0
may be written in the form
f (t) = ak ; t ∈ [k, k + 1[.
hence, V0 is the subspace of signals that are constant on intervals of the form [k, k+1[,
k ∈ Z. By exploiting the assertion (d) in Definition 2.4, we obtain for any j ∈ Z,
( )
X X
2
Vj = f ∈ L (R) ; f = aj,k ϕj,k ; (aj,k )k ⊂ R such that a2j,k <∞
k∈Z k∈Z
where we recall that

ϕj,k (t) = 2j/2 ϕ(2j t − k) = 2j/2 χ[0,1[ (2j t − k) = 2j/2 χ[k/2j ,(k+1)/2j [ (t).
Vj is by analogy the set of functions that are constant on intervals of the form
k k+1
[ j , j [, k ∈ Z.
2 2
Next, we propose to compute the coefficients of the filters H and G, and conse-
quently the expression of the associated Haar mother wavelet. The starting point is
the two-scale relation (2.22), which we recall hereafter
√ X
ϕ(x) = 2 hk ϕ(2x − k).
k
Having in hand the explicit expression of Haar scaling function ϕ = χ[0,1[ , the latter
becomes √ X
χ[0,1[ (x) = 2 hk χ[k/2,(k+1)/2[ (x).
k
Next, because of the supports of the different functions appearing in the relation, we
obtain
hk = 0 , for all k 6= 0, 1
and thus, √ √
χ[0,1[ (x) = 2h0 χ[0,1/2[ (x) + 2h1 χ[1/2,1[ (x)
which yields immediately that
1
h0 = h1 = √ .
2
It results that the filter G is given by
1
g0 = −g1 = √ and gk = 0 , for all k 6= 0, 1.
2
The associated Haar mother wavelet is then expressed by
ψ = χ[0,1/2[ − χ[1/2,1[ .
We hereafter give the graphic illustrations of the Haar scaling function ϕ and the
Haar mother wavelet ψ (see Figure 2.3).
Figure 2.3 Haar scaling and wavelet functions.
2.7.2 Faber–Schauder wavelet

It is based on the explicit Schauder scaling function
ϕ(x) = (1 − |x|)χ[−1,1[ (x)
which is also characterized by the possibility of explicit computations of the trans-
forms and coefficients. The basic approximation space V0 is composed of all functions
f ∈ L2 (R), which may be written in the form
X
f= ak ϕk
k∈Z
where, as usual, (ak )k satisfies X

a2k < ∞.
k∈Z
For k ∈ Z, ϕk is given by
ϕk (x) = ϕ0,k (x) = (1 − |x − k|)χ[−1,1[ (x − k) = (1 − |x − k|)χ[k−1,k+1[ (x).
Here, also, using the fact that the supports are disjoint, we obtain
f (x) = ak (1 − |x − k|) x ∈ [k − 1, k + 1[.
V0 is then the subspace of square integrable functions that are affine on the intervals
[k − 1, k + 1[, k ∈ Z. More precisely,
ak (1 − k + x) , x ∈ [k − 1, k[,





f (x) = ak (1 + k − x) , x ∈ [k, k + 1[,



 0 , others.
The subspace Vj is composed of square integrable functions that are of the form

j/2 j j j
 aj,k 2 (1 − k + 2 x) , x ∈ [(k − 1)/2 , k/2 [,



 j/2 j j j
f (x) = aj,k 2 (1 + k − 2 x) , x ∈ [k/2 , (k + 1)/2 [,




 0 , others.
where for all j ∈ Z, (aj,k )k satisfies

X
a2j,k < ∞.
k∈Z
We now evaluate the filter coefficients. The two-scale relation (2.22) is written in the
present case as
√ X2
(1 − |x|)χ[−1,1[ (x) = 2 hk (1 − |2x − k|)χ[(k−1)/2,(k+1)/2[ (x).
k=−2
The coefficients hk are vanishing for |k| ≥ 3. Next by choosing for x the values −1,
−1 1
, 0, and 1, we obtain
2 2
1 1
h−2 = h2 = 0 , h−1 = h1 = √ and h0 = √ .
2 2 2
The G filter will be
1 −1
g0 = g2 = √ and g1 = √ .
2 2 2
The Schauder mother wavelet will be expressed by
1 1
ψ(x) = (1 − |2x|)χ[−1/2,1/2[ (x) − (1 − |2x − 1|)χ[0,1[ (x) + (1 − |2x − 2|)χ[1/2,3/2[ (x).
2 2
The following graph (Figure 2.4) represents Schauder ϕ and ψ.
2.7.3 Daubechies wavelets

In 1988, Daubechies built with the concept of multiresolution analysis a compact
support wavelet family. These wavelets are interesting as they are orthogonal and
well localized in time or space. The specificity of this wavelet is the absence of an
explicit function. In addition, the wavelet provides the ability to choose the degree
of regularity by imposing the number of vanishing moments. We will describe the
construction of Daubechies wavelets and show some basic properties. The two basic
relationships to build a multiresolution analysis are the following:
∀ x, Φ(x) = Σ hk Φ(2x − k)
Figure 2.4 Schauder scaling and wavelet functions.
and √ X
∀ x, ψ(x) = 2 g(n)φ(2x − k)
where X
h(n − 2k)h(n − 2l) = δkl ,
n
X √
h(n) = 2,
n
and
g(n) = (−1)n h(1 − n)
If the filter H = (h(n))n∈Z is finite, the support of the scaling function is also finite.
Let
supp(φ) ⊂ [Nmin , Nmax ].
Hence,
supp(ψ) ⊂ [1 ÷ 2(1 − Nmin − Nmax ), 1 ÷ 2(1 + Nmin − Nmax )]
For example for N = 2, we obtain the so-called Db4 wavelet, whose low-pass filter
coefficients hk are √ √
1+ 3 3+ 3
h0 = √ , h1 = √
4 2 4 2
√ √
3− 3 1− 3
h2 = √ , h3 = √
4 2 4 2
The next graph illustrates some wavelet function psi due to Daubechies. The repre-
sentation illustrates clearly the strong relation between the regularity, the number of
vanishing moments and the support length N .
Figure 2.5 Some Daubechies scaling functions and associated wavelets.
2.7.4 Symlet wavelets

The symlets are nearly symmetrical wavelets proposed by Daubechies as modifica-
tions to the db family. The properties of the two wavelet families are similar. The
symlet (symN) wavelets are also known as Daubechies’ least-asymmetric wavelets. In
symN, N is the number of vanishing moments. These filters are also referred to in the
literature by the number of filter taps, which is 2N. Here are the wavelet functions
psi.
2.7.5 Spline wavelets

A fundamental remark that we can notice from the previous constructions of mul-
tiresolution analysis (and its associated scaling and wavelet functions) is the lack of
regularity. However, this fact is very important in many cases of applications such
as the resolution of partial differential equations where higher order regularities are
required. Moreover, the regularity is sometimes related to the supports of wavelets
and scaling functions as in Daubechies cases. This is also important in approximating
time series where the size is large.
In the present section, we aim to redevelop a case of explicit construction of
wavelets and thus multiresolution analysis by taking into account the characteristics
of regularity.
Figure 2.6 Some symlet scaling functions and associated wavelets.
We will start by the most simple example of wavelets and multiresolution analysis
due to the Haar case and show that, even though this is a nonregular case, it may serve
to construct higher order multiresolutions. These are the so-called spline wavelets.
So, let ϕ = χ[0,1] be the Haar scaling function, which is equivalently the char-
acteristic function of the interval [0, 1]. Let r > 0 be in N. The spline wavelets are
defined inductively by
ϕ0 = ϕ and ϕr = ϕr−1 ∗ ϕ, ∀r ≥ 1.
Definition 2.15 The function ϕr is said to be the spline wavelet of order r.

Obviously, the starting multiresolution space V0 will be the closure in L2 (R) of
spann(ϕr (. − k))k∈Z and for j ∈ W the j-level approximation space Vj may be
obtained by scaling the elements of V0 by 2j . In other words, F (.) ∈ Vj whenever
F ( 2.j ) ∈ V0 and vice versa.
We will show in the next part that spline wavelets may be seen as piecewise
polynomial functions on intervals of the form [k, k + 1[, k ∈ Z with degrees at most
the order r of the spline. More precisely, we have the following result.
Proposition 2.16 Let Pr be the space of all (r − 1) continuously differentiable func-
tions with their restriction on each interval [k, k + 1[, k ∈ Z, a polynomial of degree
at most r. Then, ϕr ∈ Pr .
Proof 2.4 We proceed by recurrence on r. The function ϕ0 is constant. Therefore, it

may be seen as a polynomial of degree 0. Assume next that on each interval [k, k + 1[,
k ∈ Z, the function ϕr is a polynomial of degree at most r. Observing next that
ϕr+1 = ϕr ∗ ϕ, it suffices to show that xr ∗ ϕ ∈ Pr . Indeed, for all x ∈ [0, 1] we
explicitly have
r
(−1)r−l l
Z 1 X
r r
(x ∗ ϕ)(t) = (y − t) dy = Crl x
0 l=0
r+1−l
which is q polynomial of degree r.
In fact, we may prove more properties of these wavelets such as the minimality
of supports. We now investigate briefly the important assumption in the construction
of multiresolution analysis stating that the collection ϕr (x − k), k ∈ Z, constitutes a
Riesz basis of the space that they span. We thus have to show that there exist constants
K1 , K2 with 0 < K1 < K2 < ∞ such that almost everywhere we have
X
K1 ≤ cr (ω + 2kπ)|2 ≤ K2 .
|ϕ
Indeed, it is straightforward that

!r+1
eiω − 1
cr (ω) =
ϕ .
iω
X
Notice now that the overlap function Γϕ = cr (ω + 2kπ)|2 is periodic with period
|ϕ
2π and it is upper bounded on [−π, π]. On the other hand, the function Θ(ω) =
exp iω − 1
| | is lower bounded away from 0 on [−π, π]. So is the function Γϕ . As a
iω
consequence, the function ϕr yields a multiresolution on R.
2.7.6 Anisotropic wavelets

The first anisotropic wavelet is due to Morlet ([279]). Its basic function is
2 /2
ψ(X) = eiK.X e−kXk
which is in fact Gaussian modulated according to the direction of K.

It is not, in any rigor, a wavelet; additional terms must be added to it for reason
of oscillation. Note that it is not isotropic but has a privileged direction given by K.
It is used to analyze images having anisotropic characteristics.
A second example of anisotropic wavelets is the so-called Mexican hat ([279]). It
is based on the function
ψ(X) = (2 − XAX)e−XAX/2
where A is an unspecified symmetric matrix of M2 (R).

It differs from that of Morlet due the fact that the anisotropy is in its module. If
A = λI, the Mexican hat is of a radial symmetry. However, if the spectrum of the
matrix A is not a singleton, the wavelet will be anisotropic.
2.7.7 Cauchy wavelets

Cauchy wavelets are one step in the direction of introducing spherical wavelets as
they aim to take into account the angular behavior of the analyzed functions. In
one-dimensional case, Cauchy wavelets are defined via their Fourier transform
(
0 for ω < 0
ψbm (ω) = m −ω
ω e for ω ≥ 0.
with m > 0. In one dimension, the positive half-line is a convex cone. Thus, a natural
generalization to two dimensions will be a wavelet whose support in spatial frequency
space is contained in a convex cone with apex at the origin. Let C ≡ C(α, β) be the
convex cone determined by the unit vectors eα , eβ , where α < β, β − α < π and for
all θ, eθ ≡ (cos θ, sin θ). The axis of the cone is ξαβ = e α+β . In other words,
2

C(α, β) = k ∈ R2 , α ≤ arg(k) ≤ β

= k ∈ R2 , k.ξαβ ≥ eα .ξαβ = eβ .> 0 .
The dual cone to C(α, β) is

C(α,
e β) = k ∈ R2 , k.k 0 > 0, ∀ k 0 ∈ C(α, β) .
Remark that C(α,

e β) may be also seen as
C(α,
e β) = C(α
b , β).
b
where
π b π
b=β−
α , β =α+
2 2
and
eα .eαb = eβ .eβb = 0.
Thus, the axis of Ce is ξαβ .

The two-dimensional Cauchy wavelet is defined via its Fourier transform
(
C,η (k.eαe)l (k.eβe)m e−k.η , k ∈ C(α, β),
ψblm (k) =
0 , otherwise.
where η ∈ Ce and l, m ∈ N∗ . Note that such a wavelet is also supported by the cone
C. It satisfies the admissibility condition
d2 k
Z
C,η
AψC,η ≡ (2π)2 |ψblm (k)|2 < ∞.
lm |k|2
The following result of Antoine et al is proved in [13] and yields an explicit form for
the two-dimensional Cauchy wavelet.
Proposition 2.17 For even η ∈ Ce and l, m ∈ N∗ , the two-dimensional Cauchy

C,η
wavelet ψlm (x) with support in C belongs to L2 (R2 , dx) and is given by
C,η il+m+2 [sin(β − α)]l+m+1

ψlm (x) = l!m! .
2π [(x + iη).eα ]l+1 [(x + iη).eβ ]m+l
We can, with analogous techniques, define multidimensional Cauchy wavelets. See

[13] and the references therein for more details.
2.8 EXERCISES
Exercise 1.
√
Denote φ = χ[0,1[ and φjk (t) = 2j φ(2j t − k), t ∈ R. Denote further E = L2 ([0, 1[)
the vector space of square-integrable functions on [0, 1[. Next for j fix V j , the vector
k k+1
space of constant functions, on { j ; j } for k ∈ {0; 2j − 1}.
2 2
a. Show that (φjk )k∈{0,2j −1} is an orthonormal basis of V j . < (φjk ); (φjh ) >= 0∀(k, h) ∈
{0, 2j − 1} × {0, 2j − 1}. and that the elements φjk are unitary.
b. Sketch the graphs of (φ1k )k∈{0,1 } and (φ2k )k∈{0,3 }.
Exercise 2.
Consider the same assumptions as in Exercise 1 above.
a. Find two functions (ψk1 )k∈{0,1} constituting an orthonormal basis of a vector space
W 1 such that V 2 = V 1 ⊕ W 1 .
b. Express the functions (ψk1 )k∈{0,1 } and (φ1k )k∈{0,1 } by means of (φ2k )k∈{0,3 }.
c. Express conversely the functions (φ2k )k∈{0,3 } in terms of (φ1k )k∈{0,1 } and
(ψk1 )k∈{0,1 }.
d. Express the function ψk1 = ψ(2t − k).
Exercise 3.
Consider the same assumptions as in Exercises 1 and 2 above.
a. Provide the 2j Haar wavelets (ψk1 )k∈{0,1 } which constitute an orthonormal basis
of W j = V j+1 V j .
b. Write the (φjk ) and (ψkj ) by means of (ψkj+1 ).
c. Write the (φj+1 j j

k ) by means of (φk ) and (ψk ).
d. Write the approximation coefficients ajk and the detail coefficients (djk ) by means
of the (aj+1
k ).
e. Write the coefficients (aj+1 j j

k ) by means of (ak ) and dk .
f. Compare the relation between the coefficients in V j ⊕ W j to the one in V j+1 .
Exercise 4.
Consider the discrete signal S = [2 4 8 12 14 0 2 1] and the Haar multiresolution
analysis on R.
a. Decomposer S in V 0 ⊕2j=0 W j
b. Sketch the graphs of S relatively to V 2 , V 1 and V 0 .
b. Sketch the graphs of S relatively to the details W 2 , W l and W 0
Exercise 5.
Let ψ : R → C be wavelet in the Schwarz class such that ψ(0)
b = 0). Denote
Z
C= |xψ(x)|dx.
R
Denote also for any function f : R → C, s > 0 and u ∈ R,

1 x−u
Z
Wf (s, u) = √ ψ f (x)dx.
s R s
1. Let f : R → C be 1-Lipschitzian. Prove that for all s and u we have
3
|Wf (s, u)| ≤ Cs 2 .
2. a) Show that there exists φ : R → C such that ψ = φ0 and ψ(x) → 0 as |x| → ±∞.
b) Let a ∈ R be fixed and consider f = 1[a,+∞] . Compute Wf by means of φ.
3. Let f : R → C be piecewise continuous with finite number of discontinuities
−
x1 < ... < xn and denote for k ≤ n αk = f (x+ k ) − f (xk ). Assume further that f
is 1-Lipschitzian on any interval not containing any of the xk ’s. Sketch the graph of
Wf (s, .) for s small enough.
Exercise 6.
Let f, ψ : R → C be in the Schwarz class such that ψ(0) = 0) and denote
Z +∞ Z
dt
F (t) = ψs (t − u)Wf (u, s)du .
0 R s2
Show that F = Cψ f , where Cψ is a constant depending on ψ and not on f .
Exercise 7.
Let ψ be a wavelet such that ψ(0)
b = 0, where
Z
ψ(w)
b = ψ(t)e−izt dt, ∀w ∈ R.
R
Another random document with
no related content on Scribd:
had not loved the children with all the devotion he should have or he
would never have had the thoughts he had had—or so he had
reasoned afterward. Yet then as now he suffered because of the love
he should have given them, and had not—and now could not any
more, save in memory. He recalled how both boys looked in those
last sad days, their pinched little faces and small weak hands! Marie
was crushed, and yet dearer for the time being than ever before. But
the two children, once gone, had seemed the victims of his own dark
thoughts as though his own angry, resentful wishes had slain them.
And so, for the time, his mood changed. He wished, if he could, that
he might undo it all, go on as before with Marie, have other children
to replace these lost ones in her affection—but no. It was apparently
not to be, not ever any more.
For, once they were gone, the cords which had held him and Marie
together were weaker, not stronger—almost broken, really. For the
charm which Marie had originally had for him had mostly been
merged in the vivacity and vitality and interest of these two prattling
curly-headed boys. Despite the financial burden, the irritation and
drain they had been at times, they had also proved a binding chain,
a touch of sweetness in the relationship, a hope for the future, a
balance which had kept even this uneven scale. With them present
he had felt that however black the situation it must endure because
of them, their growing interests; with them gone, it was rather plain
that some modification of their old state was possible—just how, for
the moment, he scarcely dared think or wish. It might be that he
could go away and study for awhile now. There was no need of his
staying here. The neighborhood was too redolent now of the
miseries they had endured. Alone somewhere else, perhaps, he
could collect his thoughts, think out a new program. If he went away
he might eventually succeed in doing better by Marie. She could
return to her parents in Philadelphia for a little while and wait for him,
working there at something as she had before until he was ready to
send for her. The heavy load of debts could wait until he was better
able to pay them. In the meantime, also, he could work and whatever
he made over and above his absolute necessities might go to her—
or to clearing off these debts.
So he had reasoned.
But it had not worked out so of course. No. In the broken mood in
which Marie then was it was not so easy. Plainly, since he had run
across her that April day in Philadelphia when he was wiring for the
great dry goods store, her whole life had become identified with his,
although his had not become merged with hers. No. She was, and
would be, as he could so plainly see, then, nothing without him,
whereas he—he—Well, it had long since been plain that he would be
better off without her—materially, anyhow. But what would she do if
he stayed away a long time—or never came back? What become?
Had he thought of that then? Yes, he had. He had even thought that
once away he might not feel like renewing this situation which had
proved so disastrous. And Marie had seemed to sense that, too. She
was so sad. True he had not thought of all these things in any bold
outright fashion then. Rather they were as sly, evasive shadows
skulking in the remote recesses of his brain, things which scarcely
dared show their faces to the light, although later, once safely away
—they had come forth boldly enough. Only at that time, and later—
even now, he could not help feeling that however much Marie might
have lacked originally, or then, the fault for their might was his,—that
if he himself had not been so dull in the first instance all these black
things would not have happened to him or to her. But could she go
on without him? Would she? he had asked himself then. And
answered that it would be better for him to leave and build himself up
in a different world, and then return and help her later. So he fretted
and reasoned.
But time had solved all that, too. In spite of the fact that he could
not help picturing her back there alone with her parents in
Philadelphia, their poor little cottage in Leigh Street in which she and
her parents had lived—not a cottage either, but a minute little brick
pigeon-hole in one of those long lines of red, treeless, smoky
barracks flanking the great mills of what was known as the
Reffington District, where her father worked—he had gone. He had
asked himself what would she be doing there? What thinking, all
alone without him—the babies dead? But he had gone.
He recalled so well the day he left her—she to go to Philadelphia,
he to Boston, presumably—the tears, the depression, the
unbelievable sadness in her soul and his. Did she suspect? Did she
foreknow? She was so gentle, even then, so trustful, so sad. “You
will come back to me, dearie, won’t you, soon?” she had said, and so
sadly. “We will be happy yet, won’t we?” she had asked between
sobs. And he had promised. Oh yes; he had done much promising in
his life, before and since. That was one of the darkest things in his
nature, his power of promising.
But had he kept that?
However much in after months and years he told himself that he
wanted to, that he must, that it was only fair, decent, right, still he
had not gone back. No. Other things had come up with the passing
of the days, weeks, months, years, other forces, other interests.
Some plan, person, desire had always intervened, interfered,
warned, counseled, delayed. Were there such counselors? There
had been times during the first year when he had written her and
sent her a little money—money he had needed badly enough
himself. Later there was that long period in which he felt that she
must be getting along well enough, being with her parents and at
work, and he had not written. A second woman had already
appeared on the scene by then as a friend. And then—
The months and years since then in which he had not done so!
After his college course—which he took up after he left Marie,
working his way—he had left Boston and gone to K—— to begin a
career as an assistant plant manager and a developer of ideas of his
own, selling the rights to such things as he invented to the great
company with which he was connected. And then it was that by
degrees the idea of a complete independence and a much greater
life had occurred to him. He found himself so strong, so interesting to
others. Why not be free, once and for all? Why not grow greater?
Why not go forward and work out all the things about which he had
dreamed? The thing from which he had extricated himself was too
confining, too narrow. It would not do to return. The old shell could
not now contain him. Despite her tenderness, Marie was not
significant enough. So—He had already seen so much that he could
do, be, new faces, a new world, women of a higher social level.
But even so, the pathetic little letters which still followed from time
to time—not addressed to him in his new world (she did not know
where he was), but to him in the old one—saying how dearly she
loved him, how she still awaited his return, that she knew he was
having a hard time, that she prayed always, and that all would come
out right yet, that they would be able to be together yet!—she was
working, saving, praying for him! True, he had the excuse that for the
first four years he had not really made anything much, but still he
might have done something for her,—might he not have?—gone
back, persuaded her to let him go, made her comfortable, brought
her somewhat nearer him even? Instead he had feared, feared,
reasoned, argued.
Yes, the then devil of his nature, his ambition, had held him
completely. He was seeing too clearly the wonder of what he might
be, and soon, what he was already becoming. Everything as he
argued then and saw now would have had to be pushed aside for
Marie, whereas what he really desired was that his great career, his
greater days, his fame, the thing he was sure to be now—should
push everything aside. And so—Perhaps he had become sharper,
colder, harder, than he had ever been, quite ready to sacrifice
everything and everybody, or nearly, until he should be the great
success he meant to be. But long before this he might have done so
much. And he had not—had not until very recently decided to revisit
this older, sweeter world.
But in the meantime, as he had long since learned, how the
tragedy of her life had been completed. All at once in those earlier
years all letters had ceased, and time slipping by—ten years really—
he had begun to grow curious. Writing back to a neighbor of hers in
Philadelphia in a disguised hand and on nameless paper, he had
learned that nearly two years before her father had died and that she
and her mother and brother had moved away, the writer could not
say where. Then, five years later, when he was becoming truly
prosperous, he had learned, through a detective agency, that she
and her mother and her ne’er-do-well brother had moved back into
this very neighborhood—this old neighborhood of his and hers!—or,
rather, a little farther out near the graveyard where their two boys
were buried. The simplicity of her! The untutored homing instinct!
But once here, according to what he had learned recently, she and
her mother had not prospered at all. They had occupied the most
minute of apartments farther out, and had finally been compelled to
work in a laundry in their efforts to get along—and he was already so
well-to-do, wealthy, really! Indeed three years before his detectives
had arrived, her mother had died, and two years after that, she
herself, of pneumonia, as had their children. Was it a message from
her that had made him worry at that time? Was that why, only six
months since, although married and rich and with two daughters by
this later marriage, he had not been able to rest until he had found
this out, returned here now to see? Did ghosts still stalk the world?
Yes, to-day he had come back here, but only to realize once and
for all now how futile this errand was, how cruel he had been, how
dreary her latter days must have been in this poor, out-of-the-way
corner where once, for a while at least, she had been happy—he
and she.
“Been happy!”
“By God,” he suddenly exclaimed, a passion of self-reproach and
memory overcoming him, “I can’t stand this! It was not right, not fair. I
should not have waited so long. I should have acted long, long since.
The cruelty—the evil! There is something cruel and evil in it all, in all
wealth, all ambition, in love of fame—too cruel. I must get out! I must
think no more—see no more.”
And hurrying to the door and down the squeaking stairs, he
walked swiftly back to the costly car that was waiting for him a few
blocks below the bridge—that car which was so representative of the
realm of so-called power and success of which he was now the
master—that realm which, for so long, had taken its meaningless
lustre from all that had here preceded it—the misery, the loneliness,
the shadow, the despair. And in it he was whirled swiftly and gloomily
away.
IX
PHANTOM GOLD
Y OU would have to have seen it to have gathered a true

impression—the stubby roughness of the country, the rocks, the
poverty of the soil, the poorness of the houses, barns, agricultural
implements, horses and cattle and even human beings, in
consequence—especially human beings, for why should they, any
more than any other product of the soil, flourish where all else was
so poor?
It was old Judge Blow who first discovered that “Jack,” or zinc,
was the real riches of Taney, if it could be said to have had any
before “Jack” was discovered. Months before the boom began he
had stood beside a smelter in far-off K—— one late winter afternoon
and examined with a great deal of care the ore which the men were
smelting, marveling at its resemblance to certain rocks or boulders
known as “slug lumps” in his home county.
“What is this stuff?” he asked of one of the bare-armed men who
came out from the blazing furnace after a time to wipe his dripping
face.
“Zinc,” returned the other, as he passed his huge, soiled palm over
his forehead.
“We have stuff down in our county that looks like that,” said the
judge as he turned the dull-looking lump over and considered for a
while. “I’m sure of it—any amount.” Then he became suddenly silent,
for a thought struck him.
“Well, if it’s really ‘Jack,’” said the workman, using the trade or
mining name for it, “there’s money in it, all right. This here comes
from St. Francis.”
The old judge thought of this for a little while and quietly turned
away. He knew where St. Francis was. If this was so valuable that
they could ship it all the way from southeast B——, why not from
Taney? Had he not many holdings in Taney?
The result was that before long a marked if secret change began
to manifest itself in Taney and regions adjacent thereto. Following
the private manipulations and goings to and fro of the judge one or
two shrewd prospectors appeared, and then after a time the whole
land was rife with them. But before that came to pass many a farmer
who had remained in ignorance of the value of his holdings was
rifled of them.
Old Bursay Queeder, farmer and local ne’er-do-well in the
agricultural line, had lived on his particular estate or farm for forty
years, and at the time that Judge Blow was thus mysteriously
proceeding to and fro and here and there upon the earth, did not
know that the rocks against which his pair of extra large feet were
being regularly and bitterly stubbed contained the very wealth of
which he had been idly and rather wistfully dreaming all his life.
Indeed, the earth was a very mysterious thing to Bursay, containing,
as it did, everything he really did not know. This collection of seventy
acres, for instance—which individually and collectively had wrung
more sweat from his brow and more curses from his lips than
anything else ever had—contained, unknown to him, the possibility
of the fulfilment of all his dreams. But he was old now and a little
queer in the head at times, having notions in regard to the Bible,
when the world would come to an end, and the like, although still
able to contend with nature, if not with man. Each day in the spring
and summer and even fall seasons he could be seen on some
portion or other of his barren acres, his stubby beard and sparse hair
standing out roughly, his fingers like a bird’s claws clutching his
plough handles, turning the thin and meagre furrows of his fields and
rattling the stony soil, which had long ceased to yield him even a
modicum of profit. It was a bare living now which he expected, and a
bare living which he received. The house, or cabin, which he
occupied with his wife and son and daughter, was dilapidated
beyond the use or even need of care. The fences were all decayed
save for those which had been built of these same impediments of
the soil which he had always considered a queer kind of stone,
useless to man or beast—a “hendrance,” as he would have said. His
barn was a mere accumulation of patchboards, shielding an old
wagon and some few scraps of machinery. And the alleged corn crib
was so aged and lopsided that it was ready to fall. Weeds and
desolation, bony horses and as bony children, stony fields and thin
trees, and withal solitude and occasional want—such was the world
of his care and his ruling.
Mrs. Queeder was a fitting mate for the life to which he was
doomed. It had come to that pass with her that the monotony of
deprivation was accepted with indifference. The absence or
remoteness of even a single modest school, meeting house or town
hall, to say nothing of convenient neighbors, had left her and hers all
but isolated. She was irascible, cantankerous, peculiar; her voice
was shrill and her appearance desolate. Queeder, whom she
understood or misunderstood thoroughly, was a source of comfort in
one way—she could “nag at him,” as he said, and if they quarreled
frequently it was in a fitting and harmonious way. Amid such a
rattletrap of fields and fences bickering was to be expected.
“Why don’t yuh take them thar slug lumps an’ make a fence over
thar?” she asked of Queeder for something like the thousandth time
in ten years, referring to as many as thirty-five piles of the best and
almost pure zinc lying along the edges of the nearest field, and piled
there by Bursay,—this time because two bony cows had invaded one
of their corn patches. The “slug lumps” to which she referred could
not have been worth less than $2,000.
For as many as the thousandth time he had replied:
“Well, fer the land sakes, hain’t I never got nuthin’ else tuh do?
Yuh’d think them thar blame-ding rocks wuz wuth more nor anythin’
else. I do well enough ez ’tis to git ’em outen the sile, I say, ‘thout
tryin’ tuh make fences outen ’em.”
“So yuh say—yuh lazy, good-fer-nuthin’ ole tobacco-chewin’ ——,”
here a long list of expletives which was usually succeeded by a
stove lid or poker or a fair-sized stick of wood, propelled by one party
or the other, and which was as deftly dodged. Love and family
affection, you see, due to unbroken and unbreakable propinquity, as
it were.
But to proceed: The hot and rainy seasons had come and gone in
monotonous succession during a period of years, and the lumps still
lay in the field. Dode, the eldest child and only son—a huge, hulking,
rugged and yet bony ignoramus, who had not inherited an especially
delicate or agreeable disposition from his harried parents—might
have removed them had he not been a “consarned lazy houn’,” his
father said, or like his father, as his mother said, and Jane, the
daughter, might have helped, but these two partook of the same
depressed indifference which characterized the father. And why not,
pray? They had worked long, had had little, seen less and hoped for
no particular outlet for their lives in the future, having sense enough
to know that if fate had been more kind there might have been.
Useless contention with an unyielding soil had done its best at
hardening their spirits.
“I don’t see no use ploughin’ the south patch,” Dode had now
remarked for the third time this spring. “The blamed thing don’t grow
nuthin’.”
“Ef yuh only half ’tended it instid o’ settin’ out thar under them thar
junipers pickin’ yer teeth an’ meditatin’, mebbe ’twould,” squeaked
Mrs. Queeder, always petulant or angry or waspish—a nature soured
by long and hopeless and useless contention.
“No use shakin’ up a lot uv rocks, ez I see,” returned Dode, wearily
and aimlessly slapping at a fly. “The hull place ain’t wuth a hill o’
beans,” and from one point of view he was right.
“Why don’t yuh git off’n hit then?” suggested Queeder in a
tantalizing voice, with no particular desire to defend the farm, merely
with an idle wish to vary the monotony. “Ef hit’s good enough tuh
s’port yuh, hit’s good enough to work on, I say.”
“S’port!” sniffed the undutiful Dode, wearily, and yet humorously
and scornfully. “I ain’t seed much s’port, ez I kin remember. Mebbe
ye’re thinkin’ uv all the fine schoolin’ I’ve had, er the places I’ve
been.” He slapped at another fly.
Old Queeder felt the sneer, but as he saw it it was scarcely his
fault. He had worked. At the same time he felt the futility of
quarreling with Dode, who was younger and stronger and no longer,
owing to many family quarrels, bearing him any filial respect. As a
matter of fact it was the other way about. From having endured many
cuffs and blows in his youth Dode was now much the more powerful
physically, and in any contest could easily outdo his father; and
Queeder, from at first having ruled and seen his word law, was now
compelled to take second, even third and fourth, place, and by
contention and all but useless snarling gain the very little
consideration that he received.
But in spite of all this they lived together indifferently. And day after
day—once Judge Blow had returned to Taney—time was bringing
nearer and nearer the tide of mining and the amazing boom that
went with it. Indeed every day, like a gathering storm cloud, it might
have been noted by the sensitive as approaching closer and closer,
only these unwitting holders were not sensitive. They had not the
slightest inkling as yet of all that was to be. Here in this roadless,
townless region how was one to know. Prospectors passed to the
north and the south of them; but as yet none had ever come directly
to this wonderful patch upon which Queeder and his family rested. It
was in too out-of-the-way a place—a briary, woodsy, rocky corner.
Then one sunny June morning—
“Hi, thar!” called Cal Arnold, their next neighbor, who lived some
three miles further on, who now halted his rickety wagon and bony
horses along the road opposite the field in which Queeder was
working. “Hyur the news?” He spoke briskly, shifting his cud of
tobacco and eyeing Queeder with the chirpiness of one who brings
diverting information.
“No; what?” asked Queeder, ceasing his “cultivating” with a worn
one-share plough and coming over and leaning on his zinc fence,
rubbing a hand through his sparse hair the while.
“Ol’ Dunk Porter down here to Newton’s sold his farm,” replied Cal,
shrewdly and jubilantly, as though he were relating the tale of a great
battle or the suspected approach of the end of the world. “An’ he got
three thousan’ dollars fer it.” He rolled the sum deliciously under his
tongue.
“Yuh don’t say!” said Queeder quietly but with profound and
amazed astonishment. “Three thousan’?” He stirred as one who
hears of the impossible being accomplished and knows it can’t be
true. “Whut fer?”
“They ’low now ez how thar’s min’l onto it,” went on the farmer
wisely. “They ’low ez how now this hyur hull kentry round hyur is
thick an’ spilin’ with it. Hit’s uvrywhar. They tell me ez how these hyur
slug lumps”—and he flicked at one of the large piles of hitherto
worthless zinc against which Queeder was leaning—“is this hyur
min’l—er ‘Jack,’ ez they call hit—an’ that hit’s wuth two cents a
pound when it’s swelted” (“smelted,” he meant) “an’ even more. I see
yuh got quite a bit uv hit. So’ve I. Thar’s a lot layin’ down around my
place. I allus ’lowed ez how ’twant wuth much o’ anythin,’ but they
say ’tis. I hyur from some o’ the boys ’at’s been to K—— that when
hit’s fixed up, swelted and like o’ that, that hit’s good fer lots of
things.”
He did not know what exactly, so he did not stop to explain.
Instead he cocked a dreamful eye, screwed up his mouth
preparatory to expectorating and looked at Queeder. The latter,
unable to adjust his thoughts to this new situation, picked up a piece
of the hitherto despised “slug” and looked at it. To think that through
all these years of toil and suffering he should have believed it
worthless and now all of a sudden it was worth two cents a pound
when “swelted” and that neighbors were beginning to sell their farms
for princely sums!—and his farm was covered with this stuff, this
gold almost! Why, there were whole hummocks of it raising slaty-
gray backs to the hot sun further on, a low wall in one place where it
rose sheer out of the ground on this “prupetty,” as he always referred
to it. Think of that! Think of that! But although he thought much he
said nothing, for in his starved and hungry brain was beginning to
sprout and flourish a great and wondrous idea. He was to have
money, wealth—ease, no less! Think of it! Not to toil and sweat in the
summer sun any more, to loaf and dream at his ease, chew all the
tobacco he wished, live in town, visit far-off, mysterious K——, see
all there was to see!
“Well, I guess I’ll be drivin’ on,” commented Arnold after a time,
noting Queeder’s marked abstraction. “I cal’late tuh git over tuh
Bruder’s an’ back by sundown. He’s got a little hay I traded him a pig
fer hyur a while back,” and he flicked his two bony horses and was
off up the rubbly, dusty road.
For a time Queeder was scarcely satisfied to believe his senses.
Was it really true? Had Porter really sold his place? For days
thereafter, although he drove to Arno—sixteen miles away—to
discover the real truth, he held his own counsel, nursing a wonderful
fancy. This property was his, not his wife’s, nor his two children’s.
Years before he had worked and paid for it, a few lone dollars at a
time, or their equivalent in corn, pigs, wheat, before he had married.
Now—now—soon one of those strange creatures—a “prowspector,”
Arnold had called him—who went about with money would come
along and buy up his property. Wonderful! Wonderful! What would he
get for it?—surely five thousand dollars, considering that Porter had
received three for forty acres, whereas he had seventy. Four
thousand, anyhow—a little more than Dunk. He could not figure it
very well, but it would be more than Dunk’s, whatever it was—
probably five thousand!
The one flaw in all this though—and it was a great flaw—was the
thought of his savage and unkindly family—the recalcitrant Dode, the
angular Jane and his sour better half, Emma—who would now
probably have to share in all this marvelous prosperity, might even
take it away from him and push him into that background where he
had been for so long. They were so much more dogmatic, forceful
than he. He was getting old, feeble even, from long years of toil. His
wife had done little this long time but sneer and jeer at him, as he
now chose most emphatically to remember; his savage son the
same. Jane, the indifferent, who looked on him as a failure and a
ne’er-do-well, had done nothing but suggest that he work harder.
Love, family tenderness, family unity—if these had ever existed they
had long since withered in the thin, unnourishing air of this rough,
poverty-stricken world. What did he owe any of them? Nothing. And
now they would want to share in all this, of course. Having lived so
long with them, and under such disagreeable conditions, he now
wondered how they would dare suggest as much, and still he knew
they would. Fight him, nag him, that’s all they had ever done. But
now that wealth was at his door they would be running after him,
fawning upon him—demanding it of him, perhaps! What should he
do? How arrange for all of this?—for wealth was surely close to his
hands. It must be. Like a small, half-intelligent rat he peeked and
perked. His demeanor changed to such an extent that even his
family noticed it and began to wonder, although (knowing nothing of
all that had transpired as yet) they laid it to the increasing queerness
of age.
“Have yuh noticed how Pap acts these hyur days?” Dode inquired
of Jane and his mother one noontime after old Queeder had eaten
and returned to the fields. “He’s all the time standin’ out thar at the
fence lookin’ aroun’ ez if he wuz a-waitin’ fer somebody er thinkin’
about somepin. Mebbe he’s gittin’ a little queer, huh? Y’ think so?”
Dode was most interested in anything which concerned his father
—or, rather, his physical or mental future—for once he died this
place would have to be divided or he be called upon to run it, and in
that case he would be a fitting catch for any neighborhood farming
maiden, and as such able to broach and carry through the long-
cherished dream of matrimony, now attenuated and made all but
impossible by the grinding necessities he was compelled to endure.
“Yes, I’ve been noticin’ somepin,” returned Mrs. Queeder. “He
hain’t the same ez he wuz a little while back. Some new notion he’s
got into his mind, I reckon, somepin he wants tuh do an’ kain’t, er
somepin new in ’ligion, mebbe. Yuh kain’t ever tell whut’s botherin’
him.”
Jane “’lowed” as much and the conversation ended. But still
Queeder brooded, trying to solve the knotty problem, which
depended, of course, on the open or secret sale of the land—secret,
if possible, he now finally decided, seeing that his family had always
been so unkind to him. They deserved nothing better. It was his—
why not?
In due time appeared a prospector, mounted on horseback and
dressed for rough travel, who, looking over the fields of this area and
noting the value of these particular acres, the surface outcropping of
a thick vein, became intensely interested. Queeder was not to be
seen at the time, having gone to some remote portion of the farm,
but Mrs. Queeder, wholly ignorant of the value of the land and
therefore of the half-suppressed light in the stranger’s eye, greeted
him pleasantly enough.
“Would you let me have a drink of water?” inquired the stranger
when she appeared at the door.
“Sartinly,” she replied with a tone of great respect. Even
comparatively well-dressed strangers were so rare here.
Old Queeder in a distant field observing him at the well, now
started for the house.
“What is that stuff you make your fences out of?” asked the
stranger agreeably, wondering if they knew.
“Well, now, I dunno,” said Mrs. Queeder. “It’s some kind o’ stone, I
reckon—slug lumps, we uns always call hit aroun’ hyur.”
The newcomer suppressed a desire to smile and stooped to pick
up a piece of the zinc with which the ground was scattered. It was
the same as he had seen some miles back, only purer and present
in much greater quantities. Never had he seen more and better zinc
near the surface. It was lying everywhere exposed, cultivation, frosts
and rains having denuded it, whereas in the next county other men
were digging for it. The sight of these dilapidated holdings, the
miserable clothing, old Queeder toiling out in the hot fields, and all
this land valueless for agriculture because of its wondrous mineral
wealth, was almost too much for him.
“Do you own all this land about here?” he inquired.
“’Bout seventy acres,” returned Mrs. Queeder.
“Do you know what it sells at an acre?”
“No. It ain’t wuth much, though, I reckon. I ain’t heerd o’ none bein’
sold aroun’ hyur fer some time now.”
The prospector involuntarily twitched at the words “not wuth
much.” What would some of his friends and rivals say if they knew of
this particular spot? What if some one should tell these people? If he
could buy it now for a song, as he well might! Already other
prospectors were in the neighborhood. Had he not eaten at the same
table at Arno with three whom he suspected as such? He must get
this, and get it now.
“I guess I’ll stroll over and talk to your husband a moment,” he
remarked and ambled off, the while Mrs. Queeder and Jane, the
twain in loose blue gingham bags of dresses much blown by the
wind, stood in the tumbledown doorway and looked after him.
“Funny, ain’t he?” said the daughter. “Wonder whut he wants o’
Paw?”
Old Queeder looked up quizzically from his ploughing, to which he
had returned, as he saw the stranger approaching, and now
surveyed him doubtfully as he offered a cheery “Good morning.”
“Do you happen to know if there is any really good farming land
around here for sale?” inquired the prospector after a few delaying
comments about the weather.
“Air yuh wantin’ it fer farmin’?” replied Queeder cynically and
casting a searching look upon the newcomer, who saw at once by
Queeder’s eye that he knew more than his wife. “They’re buyin’ hit
now mostly fer the min’l ez is onto it, ez I hyur.” At the same time he
perked like a bird to see how this thrust had been received.
The prospector smiled archly if wisely. “I see,” he said. “You think
it’s good for mining, do you? What would you hold your land at as
mineral land then if you had a chance to sell it?”
Queeder thought for a while. Two wood doves cooed mournfully in
the distance and a blackbird squeaked rustily before he answered.
“I dunno ez I keer tuh sell yit.” He had been getting notions of late
as to what might be done if he were to retain his land, bid it up
against the desires of one and another, only also the thought of how
his wife and children might soon learn and insist on dividing the
profits with him if he did sell it was haunting him. Those dreams of
getting out in the world and seeing something, of getting away from
his family and being happy in some weird, free way, were actually
torturing him.
“Who owns the land just below here, then?” asked the stranger,
realizing that his idea of buying for little or nothing might as well be
abandoned. But at this Queeder winced. For after all, the land
adjoining had considerable mineral on it also, as he well knew.
“Why, let me see,” he replied waspishly, with mingled feelings of
opposition and indifference. “Marradew,” he finally added, grudgingly.
It was no doubt true that this stranger or some other could buy of
other farmers if he refused to sell. Still, land around here anywhere
must be worth something, his as much as any other. If Dunk Porter
had received $3,000—
“If you don’t want to sell, I suppose he might,” the prospector
continued pleasantly. The idea was expressed softly, meditatively,
indifferently almost.
There was a silence, in which Queeder calmly leaned on his
plough handles thinking. The possibility of losing this long-awaited
opportunity was dreadful. But he was not floored yet, for all his
hunger and greed. Arnold had said that the metal alone, these rocks,
was worth two cents a pound, and he could not get it out of his mind
that somehow the land itself, the space of soil aside from the metal,
must be worth something. How could it be otherwise? Small crops of
sorts grew on it.
“I dunno,” he replied defiantly, if internally weakly. “Yuh might ast. I
ain’t heerd o’ his wantin’ to sell.” He was determined to risk this last if
he had to run after the stranger afterward and beg him to
compromise, although he hoped not to have to do that, either. There
were other prospectors.
“I don’t know yet whether I want this,” continued the prospector
heavily and with an air of profound indifference, “but I’d like to have
an option on it, if you’d like to sell. What’ll you take for an option at
sixty days on the entire seventy acres?”
The worn farmer did not in the least understand what was meant
by the word option, but he was determined not to admit it. “Whut’ll
yuh give?” he asked finally, in great doubt as to what to say.
“Well, how about $200 down and $5,000 more at the end of sixty
days if we come to terms at the end of that time?” He was offering
the very lowest figure that he imagined Queeder would take, if any,
for he had heard of other sales in this vicinity this very day.
Queeder, not knowing what an option was, knew not what to say.
Five thousand was what he had originally supposed he might be
offered, but sixty days! What did he mean by that? Why not at once if
he wanted the place—cash—as Dunk Porter, according to Arnold,
had received? He eyed the stranger feverishly, fidgeted with his
plough handles, and finally observed almost aimlessly: “I ’low ez I
could git seven thousan’ any day ef I wanted to wait. The feller hyur
b’low me a ways got three thousan’, an’ he’s got thirty acres less’n I
got. Thar’s been a feller aroun’ hyur offerin’ me six thousan’.”
“Well, I might give you $6,000, providing I found the ground all
right,” he said.
“Cash down?” asked Queeder amazedly, kicking at a clod.
“Within sixty days,” answered the prospector.
“Oh!” said Queeder, gloomily. “I thort yuh wanted tuh buy t’day.”
“Oh, no,” said the other. “I said an option. If we come to terms I’ll
be back here with the money within sixty days or before, and we’ll
close the thing up—six thousand in cash, minus the option money.
Of course I don’t bind myself absolutely to buy—just get the privilege
of buying at any time within sixty days, and if I don’t come back
within that time the money I turn over to you to-day is yours, see,
and you’re free to sell the land to some one else.”
“Huh!” grunted Queeder. He had dreamed of getting the money at
once and making off all by himself, but here was this talk of sixty
days, which might mean something or nothing.
“Well,” said the prospector, noting Queeder’s dissatisfaction and
deciding that he must do something to make the deal seem more
attractive, “suppose we say seven thousand, then, and I put down
$500 cash into your hands now? How’s that? Seven thousand in
sixty days and five hundred in cash right now. What do you say?”
He reached in his pocket and extracted a wallet thick with bills,
which excited Queeder greatly. Never had so much ready money,
which he might quickly count as his if he chose, been so near him.
After all, $500 in cash was an amazing amount in itself. With that
alone what could he not do? And then the remainder of the seven
thousand within sixty days! Only, there were his wife and two
children to consider. If he was to carry out his dream of decamping
there must be great secrecy. If they learned of this—his possession
of even so much as five hundred in cash—what might not happen?
Would not Dode or his wife or Jane, or all three, take it away from
him—steal it while he was asleep? It might well be so. He was so
silent and puzzled that the stranger felt that he was going to reject
his offer.
“I’ll tell you what I’ll do,” he said, as though he were making a
grand concession. “I’ll make it eight thousand and put up eight
hundred. How’s that? If we can’t arrange it on that basis we’ll have to
drop the matter, for I can’t offer to pay any more,” and at that he
returned the wallet to his pocket.
But Queeder still gazed, made all but dumb by his good fortune
and the difficulties it presented. Eight thousand! Eight hundred in
cash down! He could scarcely understand.
“T’day?” he asked.
“Yes, to-day—only you’ll have to come with me to Arno. I want to
look into your title. Maybe you have a deed, though—have you?”
Queeder nodded.
“Well, if it’s all right I’ll pay you the money at once. I have a form of
agreement here and we can get some one to witness it, I suppose.
Only we’ll have to get your wife to sign, too.”
Queeder’s face fell. Here was the rub—his wife and two children!
“She’s gotta sign, hez she?” he inquired grimly, sadly even. He was
beside himself with despair, disgust. To work and slave so all these
years! Then, when a chance came, to have it all come to nothing, or
nearly so!
“Yes,” said the prospector, who saw by his manner and tone that
his wife’s knowledge of it was not desired. “We’ll have to get her
signature, too. I’m sorry if it annoys you, but the law compels it.
Perhaps you could arrange all that between you in some way. Why
not go over and talk to her about it?”
Queeder hesitated. How he hated it—this sharing with his wife and
son! He didn’t mind Jane so much. But now if they heard of it they
would quarrel with him and want the larger share. He would have to
fight—stand by his “rights.” And once he had the money—if he ever
got it—he would have to watch it, hide it, to keep it away from them.
“What’s the matter?” asked the prospector, noting his perturbation.
“Does she object to your selling?”
“’Tain’t that. She’ll sell, well enough, once she hyurs. I didn’t ’low
ez I’d let ’er know at fust. She’ll be wantin’ the most uv it—her an’
Dode—an’ hit ain’t ther’n, hit’s mine. I wuz on hyur fust. I owned this
hyur place fust, ’fore ever I saw ’er. She don’t do nuthin’ but fuss an’
fight, ez ’tis.”
“Supposing we go over to the house and talk to her. She may not
be unreasonable. She’s only entitled to a third, you know, if you don’t
want to give her more than that. That’s the law. That would leave you
nearly five thousand. In fact, if you want it, I’ll see that you get five
thousand whatever she gets.” He had somehow gathered the
impression that five thousand, for himself, meant a great deal to
Queeder.
And true enough, at that the old farmer brightened a little. For five
thousand? Was not that really more than he had expected to get for
the place as a whole but an hour before! And supposing his wife did
get three thousand? What of it? Was not his own dream coming
true? He agreed at once and decided to accompany the prospector
to the house. But on the way the farmer paused and gazed about
him. He was as one who scarcely knew what he was doing. All this
money—this new order of things—if it went through! He felt strange,
different, confused. The mental ills of his many years plus this great
fortune with its complications and possibilities were almost too much
for him. The stranger noted a queer metallic and vacant light in the
old farmer’s eyes as he now turned slowly about from west to east,
staring.
“What’s the matter?” he asked, a suspicion of insanity coming to
him.
The old man seemed suddenly to come to. “’Tain’t nuthin’,” he
said. “I wuz just thinkin’.”
The prospector meditated on the validity of a contract made with a
lunatic, but the land was too valuable to bother about trifles. Once a
contract was made, even with a half-wit, the legal difficulties which
could be made over any attempt to break the agreement would be
very great.
In the old cabin Jane and her mother wondered at the meaning of
the approaching couple, but old Queeder shooed off the former as
he would have a chicken. Once inside the single room, which served
as parlor, sitting-room, bedroom and all else convenient, Queeder
nervously closed the door leading into the kitchen, where Jane had
retired.
“Go on away, now,” he mumbled, as he saw her there hanging
about. “We want a word with yer Maw, I tell yuh.”
Lank Jane retired, but later clapped a misshapen ear to the door
until she was driven away by her suspicious father. Then the farmer
began to explain to his wife what it was all about.
“This hyur stranger—I don’t know your name yit—”
“Crawford! Crawford!” put in the prospector.
“Crawford—Mr. Crawford—is hyur tuh buy the place ef he kin. I
thought, seein’ ez how yuh’ve got a little int’est in it—third”—he was
careful to add—“we’d better come an’ talk tuh yuh.”
“Int’est!” snapped Mrs. Queeder, sharply and suspiciously, no
thought of the presence of the stranger troubling her in her
expression of her opinion, “I should think I had—workin’ an’ slavin’

Wavelet Analysis Basic Concepts and Applications 1St Edition Sabrine Arfaoui Online Ebook Texxtbook Full Chapter PDF

Uploaded by

Copyright:

Available Formats

You might also like

Wavelet Analysis Basic Concepts and Applications 1St Edition Sabrine Arfaoui Online Ebook Texxtbook Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wavelet Analysis Basic Concepts and Applications 1St Edition Sabrine Arfaoui Online Ebook Texxtbook Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Wavelet Analysis: Basic Concepts and

Applications 1st Edition Sabrine

Environmental Biotechnology Basic Concepts and

Introductory Immunology: Basic Concepts for

Pharmacokinetic and Pharmacodynamic Data Analysis

Big Data Analysis for Green Computing: Concepts and

Primary Mathematics 3A Hoerst

Heterophase Polymerization Basic Concepts and

Hydrochemistry Basic Concepts and Exercises 2nd Edition

Rowland and Tozer s Clinical Pharmacokinetics and

Anouar Ben Mabrouk

and by CRC Press

Library of Congress Cataloging-in-Publication Data

Names: Arfaoui, Sabrine, author. | Ben Mabrouk, Anouar, author. | Cattani,

ISBN: 978-0-367-56218-2 (hbk)

Typeset in Latin Modern font

Chapter 2 Wavelets on Euclidean Spaces 5

Chapter 3 Wavelets extended 35

3.1 AFFINE GROUP WAVELETS 35

3.2.1 Monasse–Perrier construction 37

Chapter 4 Clifford wavelets 51

Chapter 5 Quantum wavelets 99

5.6 SOME QUANTUM SPECIAL FUNCTIONS 123

Chapter 6 Wavelets in statistics 137

6.1 INTRODUCTION 137

Chapter 7 Wavelets for partial differential equations 163

7.1 INTRODUCTION 163

7.3 WAVELET GALERKIN APPROACH 166

Chapter 8 Wavelets for fractal and multifractal functions 183

8.1 INTRODUCTION 183

2.1 Decomposition algorithm. 15

4.1 Wedge product of vectors. 58

5.1 Graphs of Bessel functions Jv of the first kind for v = 0, 1, 2, 3. 102

8.1 Wavelet decomposition of the SP500 signal at the level J = 4. 203

Wavelets were discovered in the eighteenth century, essentially in pertroleum extrac-

Organization of the book

Chapter 3 is a mixture of different concepts. We aim precisely to develop it as a

• Localization in time/frequency domains

• Enough vanishing moments,

To analyze a signal by wavelets, one passes as in Fourier analysis by the wavelet

2.2.1 Continuous wavelet transform

Sometimes, we say that ψ is C m on R. The time localization chart is a normalization

1. The linearity, in the sense that

ds,u (αf + βg) = αds,u (f ) + βds,u (g), ∀f, g.

2. The translation-invariance, in the sense that

ds,u (τt f ) = ds,u−t (f ), ∀f ; and ∀u, s, t,

3. The dilation-invariance, in the sense that

and where for a > 0,

Theorem 2.2 For all f ∈ L2 (R), we have the L2 -equality

The proof of this result is based on the following lemma:

Lemma 2.3 Under the hypothesis of Theorem (2.2), we have

By application of Fubini’s rule, we get

Proof of Theorem 2.2. By applying the Riesz rule, we get

which by Cauchy–Schwarz inequality is bounded by

2.2.2 Discrete wavelet transform

2.3 MULTI-RESOLUTION ANALYSIS

Definition 2.4 A multiresolution analysis is a sequence of closed subsets (Vj )j∈Z of

c) There exists ϕ ∈ V0 such that {ϕ0,k = ϕ(. − k); k ∈ Z} is a Riesz basis of V0 .

where the SWj for j ∈ Z designates the orthogonal