Wavelet Analysis Basic Concepts And Applications

1st Edition Sabrine Arfaoui download

https://ebookbell.com/product/wavelet-analysis-basic-concepts-
and-applications-1st-edition-sabrine-arfaoui-51684912

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Wavelet Analysis On The Sphere Spheroidal Wavelets Sabrine Arfaoui

Imen Rezgui Anouar Ben Mabrouk Knowledge Unlatched

https://ebookbell.com/product/wavelet-analysis-on-the-sphere-
spheroidal-wavelets-sabrine-arfaoui-imen-rezgui-anouar-ben-mabrouk-
knowledge-unlatched-51110958

Wavelet Analysis And Transient Signal Processing Applications For

Power Systems He

https://ebookbell.com/product/wavelet-analysis-and-transient-signal-
processing-applications-for-power-systems-he-22035766

Wavelet Analysis Twenty Years Developments Proceedings Of The

International Conference Of Computational Harmonic Analysis Hong Kong
China 48 June 2001 International Conference Of Computational Harmonic
Analysis
https://ebookbell.com/product/wavelet-analysis-twenty-years-
developments-proceedings-of-the-international-conference-of-
computational-harmonic-analysis-hong-kong-
china-48-june-2001-international-conference-of-computational-harmonic-
analysis-2632376

Wavelet Analysis On Local Fields Of Positive Characteristic

Biswaranjan Behera

https://ebookbell.com/product/wavelet-analysis-on-local-fields-of-
positive-characteristic-biswaranjan-behera-37754402
Wavelet Analysis In Civil Engineering Pranesh Chatterjee

https://ebookbell.com/product/wavelet-analysis-in-civil-engineering-
pranesh-chatterjee-4986866

Wavelet Analysis In Civil Engineering Pranesh Chatterjee

https://ebookbell.com/product/wavelet-analysis-in-civil-engineering-
pranesh-chatterjee-5059264

Wavelet Analysis And Its Applications Second International Conference

Waa 2001 Hong Kong China December 1820 2001 Proceedings 1st Edition
John Daugman Auth

https://ebookbell.com/product/wavelet-analysis-and-its-applications-
second-international-conference-waa-2001-hong-kong-china-
december-1820-2001-proceedings-1st-edition-john-daugman-auth-1147700

Wavelet Analysis And Applications 1st Edition Tao Qian Tao Qianmang I
Vaixu Yuesheng

https://ebookbell.com/product/wavelet-analysis-and-applications-1st-
edition-tao-qian-tao-qianmang-i-vaixu-yuesheng-1189502

Wavelet Analysis For Robust Speech Processing And Applications

Applications Of Discrete Wavelet Transform And Wavelet Denoising To
Speech Classification Enhancement And Robust Speech Recognition Tuan
Van Pham
https://ebookbell.com/product/wavelet-analysis-for-robust-speech-
processing-and-applications-applications-of-discrete-wavelet-
transform-and-wavelet-denoising-to-speech-classification-enhancement-
and-robust-speech-recognition-tuan-van-pham-1441648
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 [email protected]

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

7.1 INTRODUCTION 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

8.1 INTRODUCTION 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.
4  Wavelet Analysis: Basic Concepts and Applications

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
6  Wavelet Analysis: Basic Concepts and Applications

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|
8  Wavelet Analysis: Basic Concepts and Applications

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.

 Wavelets on Euclidean Spaces  9

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
10  Wavelet Analysis: Basic Concepts and Applications

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.
 Wavelets on Euclidean Spaces  11

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
12  Wavelet Analysis: Basic Concepts and Applications

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 >
−∞
 Wavelets on Euclidean Spaces  13

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
14  Wavelet Analysis: Basic Concepts and Applications

• 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
 Wavelets on Euclidean Spaces  15

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
16  Wavelet Analysis: Basic Concepts and Applications

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
 Wavelets on Euclidean Spaces  17

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|
18  Wavelet Analysis: Basic Concepts and Applications

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
 Wavelets on Euclidean Spaces  19

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 ).
20  Wavelet Analysis: Basic Concepts and Applications

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
 Wavelets on Euclidean Spaces  21

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 .
22  Wavelet Analysis: Basic Concepts and Applications

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
 Wavelets on Euclidean Spaces  23

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).
24  Wavelet Analysis: Basic Concepts and Applications

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.
 Wavelets on Euclidean Spaces  25

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)
26  Wavelet Analysis: Basic Concepts and Applications

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 .
 Wavelets on Euclidean Spaces  27

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.
28  Wavelet Analysis: Basic Concepts and Applications

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 .
 Wavelets on Euclidean Spaces  29

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.
30  Wavelet Analysis: Basic Concepts and Applications

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.
 Wavelets on Euclidean Spaces  31

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 ).
32  Wavelet Analysis: Basic Concepts and Applications

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
 Wavelets on Euclidean Spaces  33

Show the following Heisenberg uncertainty inequality:

2
kψk22 ≤ √ ktψ(t)k2 kξ ψ(ξ)k
b 2.
2π
(Hint. We may use the derivative (tψ(t))0 ).

Exercise 8.
Let f ∈ L2 (R) be such that fb(ξ) = 0 for ξ < 0. For z ∈ C such that I, (z) ≥ 0, let

1
Z +∞
fp (z) = (iξ)p fb(ξ)eizξ dξ.
π 0

a) Prove that fp is the derivative of f at the order p whenever f is C p .

b) Prove that fp is analytic on the upper half complex plane Im(z) > 0.
c) Prove that for z = x + iy, y > 0, we have

fp (z) = y −p−1/2 dy,x (f ),

calculated with an analytic wavelet ψ that you will specify.

Exercise 9.
Let for N ∈ N, f : [0, N ] → R be continuous and (Vm )m≥0 be a multiresolution
analysis on [0, N ] with a scaling function ϕ.
1) Verify that for all ε > 0 there exists m ≥ 0 such that for all elements t1 , t2 ∈ [0, N ]
with |t1 − t2 | < 2−m we have |f (t1 ) − f (t2 )| < ε.
2) Denote next tm,n+1/2 = (n + 21 )2−m and
N −1
2X
fε (t) = f (tm,n+1/2 )ϕm,n (t).
n=0

Prove that
|f (t) − fε (t)| < ε, ∀t ∈ [n2−m , (n + 1)2−m ].
3) Show that √
kf − fε k2 ≤ ε N .

Exercise 10.
Part A. Consider the Haar system on R with its scaling function ϕ = χ[0,1[ and
wavelet function ψ(x) = χ[0,1/2[ − χ[1/2,1[ . Let next f (t) = (t2 + 1)χ[−1,1[ and its
projections on Vj and Wj be, respectively,
X X
fj = aj,k ϕj,k and dj = dj,k ψj,k .
k k

a) Evaluate f0 (t), f1 (t), f2 (t), d0 (t) and d1 (t).

b) Draw the graphs of f (t), f0 (t), f1 (t), f2 (t), d0 (t) and d1 (t).
34  Wavelet Analysis: Basic Concepts and Applications

Part B. Denote for (x, y) ∈ R2

φ(x, y) = ϕ(x)ϕ(y), Ψ0 (x, y) = ϕ(x)ψ(y),

Ψ1 (x, y) = ψ(x)ϕ(y), Ψ2 (x, y) = ψ(x)ψ(y).

a) Prove that the system (Φ, Ψi , i = 0, 1, 2) constitutes a scaling function and a
wavelet on R2 .
b) Express the approximation and detail spaces Vj and Wj associated.
c) Let f (x, y) = cos(x) sin(x+y) on the cube C = [−π, π]×[−π, π] and its projections
on Vj and Wj be, respectively,
X X
fj = aj,k Φj,k and dij = dj,k Ψij,k , i = 0, 1, 2.
k k

a) Evaluate f0 (x, y), f1 (x, y), f2 (x, y), d0 (x, y) and d1 (x, y).
b) Draw the graphs of f (x, y),f0 (x, y), f1 (x, y), f2 (x, y), d0 (x, y) and d1 (x, y).
 CHAPTER 3

Wavelets extended

3.1 AFFINE GROUP WAVELETS

Wavelet transform of functions may be understood as a group action on the space
of square integrable functions L2 (R) for suitable group. In this chapter, we aim to
present the eventual link between wavelet transforms and the real affine group. To
make the section self-contained and easy for readers from different fields, we firstly
recall the basic definition of groups.

Definition 3.1 Let G be a nonempty set. G is said to be a group if there exists an

operation ∗ defined on G for which

1. x ∗ y ∈ G for all x, y ∈ G (known as closure property).

2. (z ∗ y) ∗ z = z ∗ (y ∗ 2) for all x, y, z ∈ G (known as associative property).

3. There is an element e ∈ G such that e ∗ x = x ∗ e = x for all x ∈ G (e is known

as the identity element).

4. For each x ∈ G, there is an element x

e ∈ G such that x ∗ x e ∗ x = e (x
e=x e is
called the inverse of x).

If, in addition, x ∗ y = y ∗ x for all x, y ∈ G, the group G is said to be commutative.

A subset H of a group G is a subgroup of G if it is a group under the same operation

as G.
The group G may have additional structure. For instance, if G is a topological
space and the operation ∗ is continuous on G × G with respect to this topology, G
will be a topological group. Furthermore, if G has a differentiable structure, G will
be a differentiable manifold, and is called a Lie group. A differentiable manifold, for
our purpose, is a set that looks like Euclidean space. For example, Rn is one since it
is Euclidean space. On Euclidean spaces, we can choose different parametrizations;
however, the change of coordinates must be differentiable. Rectangular coordinates
and polar coordinates are familiar examples of such parametrizations of Rn . We recall
that change of coordinates must also be invertible locally, that is, the Jacobian of the
transformation should be invertible.

35
36  Wavelet Analysis: Basic Concepts and Applications

In our case, we will focus on the so-called linear or affine group on Rn . These
are groups that can be realized as groups of linear transformations of a vector space
or, equivalently, as matrix groups. We will see that group representation theory may
help explain the essential properties of wavelet transforms.
Consider the connected affine group G+ , also called the ax + b group, consisting
of transformations of R of the type x 7−→ gx = ax + b, x ∈ R, where a > 0, b ∈ R
and where we denoted g = (b, a) ∈ G+ with the multiplication law

gl g2 = (b1 + a1 b2 , a1 a2 ).

dadb
It holds that an invariant measure on G is dµ (a, b) = .
a2
A nontrivial unitary irreducible representations of G on Hilbert space L2 (R, dx)
will be
1 x−b
(U (a, b)f )(x) = √ f ( ), f ∈ L2 (R, dx), (a, b) ∈ G.
a a
For the sequel, we define the following operators on the set of functions defined on
the Euclidean space R:

• The translation: Tb f (x) = f (x − b), a ∈ R.

√
• The dilation: Da f (x) = af (ax), a > 0.

• The modulation: Ea f (x) = e2iπax f (x), a ∈ R.

• The multiplication: Ma f (x) = e2iπa f (x), a ∈ R.

These operators are easily seen to be unitary. Using these operators, we immediately
deduce that
(U (a, b)f ) = Da−1 Tb f.
As a consequence, the wavelet transform of f at the scale a > 0 and the position
b ∈ R may be expressed by means of the representation U as

Cf (a, b) = hf, U (a, b)ψi = hf, Da−1 Tb ψi.

This inverse transform will be expressed as

1
f (•) = hCf (a, b), U (a, b)ψ(•)idµ (a,b) ,
Aψ

where Aψ is the admissibility constant expressed as

2
|ψ(ξ)|
Z b
Aψ = .
R |ξ|

The results presented in this section may be extended in a usual way to multidimen-
sional case.
 Wavelets extended  37

3.2 MULTIRESOLUTION ANALYSIS ON THE INTERVAL

Different ways have been proposed to adapt the AMR of L2 (R) to L2 (]0, 1[). [68,
69, 70, 71, 72]. Most of these constructions are based on the idea of considering the
scaling functions ϕj,k with supports in [0, 1] and adding suitable linear combinations
of elements ϕj,k whose support crosses the left (respect. the right) boundary of the
interval.

3.2.1 Monasse–Perrier c onstruction

Monasse and Perrier [304] have proceeded as follows: for all integer j ≥ j0 = log(4N ),
]0,1[
the space Vj is spanned by the following functions.

• 2j/2 ϕ̃0k (2j x)|]0,1[ , k = 0, · · · , M,

• ϕj,k K = N, · · · , 2j − N,

• 2j/2 ϕ̃1k (2j (x − 1))|]0,1[ , k = 0, · · · , M,

where the functions ϕ̃0k and ϕ̃1k are defined as linear combinations of ϕj,k with supports
containing 0 or 1 and with suitable polynomial coefficients such as
N
X −1
• ϕ̃0k (x) = Pk0 (l)ϕ(x − l),
−N +1

N
X −1
• ϕ̃1k (x) = Pk1 (l)ϕ(x − l).
−N +1

P00 , P10 , ..., PM

0
and P01 , P11 , ..., PM
1
are two bases of the space of polynomials of degree
at most M . The functions ϕj,k , k = N, ..., 2j − N , whose supports are contained in
[0,1], are called interior scaling functions. The functions ϕ̃0k and ϕ̃1k , k = 0, ..., M are
called edge scaling functions.

3.2.2 Bertoluzza–Falletta construction

Recently, Bertoluzza and Faletta [70] have proposed a new construction of AMR on
the interval by considering the following spaces

Vjloc = Span < ϕj,k , k ∈ Z >L2loc (R) ,

where ϕ is the scaling function of an AMR in L2 (R) with compact support. Let Vj∗
be the subspace of Vjloc of functions with coefficients having polynomial behavior on
some neighborhood of 0 or 1. The degrees of polynomials depend possibly on j. More
precisely, they have proceeded as follows. For j ≥ 0, let Nj = 2j − 2N + 2M + 3, Mj =
min(M, Nj − 1) and
( )
fkj ϕj,k ; ∃Pl , Pr PMj , fkj 1, fkj
X
Vj∗ = ∈ = Pl (k), k ≤ N − = Pr (k), k ≥ 2j − N + 1 .
k∈Z
Random documents with unrelated
content Scribd suggests to you:
Thalásio, a Seawater-colour or blew like the Sea-waues.
Thalassagliéne, an hearbe very salt in taste.
Thalasséglie, as Potamánti.
Thalassiárca, an Admirall.
Thalassíno, as Thalásio.
Thalassinténo, a kind of salt wine vsed among the Grecians.
Thalassoméli, a kind of surrop made of salt sea-water.
Thália, one of the nine Muses, as much to say capacity.
Thaliétro, the hearbe Laske-wort or Flax-wort.
Thállo, vsed of the Florentines for the colour Greene. Vsed also for a
bough that still continueth greene.
Thalmudísta, a professor of the Chaldeans or Egiptians old lawes.
Thalofóri, bringers of branches or boughes.

Thápsia, an hearbe like Dill, which being cut yeeldes a iuice as white as
milke.
Thargélie fẻste, certaine feastes and games kept in honour of Apóllo.
Thásia, a kind of wine in Egypt. Also as Thápsia.
Tháspi, a kind of Mustard-seed.
Tháu, an Hebrew letter vsed often misteriously for the crosse of Christ, as
also for the number of 400.
Theamẻde, a stone which contrary to the Adamant putteth away iron from
it.
Theangélide, a kind of hearb.
Theatineríe, hipocriticall fryers trikes.
Theatíni, a kind of order of friars, that much dissemble and faine purity.
Theatrále, of or belonging to a Theater.
Theátrico, a plaier in a Theater, a Stage plaier.
Theátro, a Theater, an open place where all men may sit, see and behold,
or wherein one sheweth, plaieth or declareth. Also a stage or skaffold.
Also any spectacle or common-play.
Thebáide, a kind of poore drie Date.
Theicóne, as Trogéte.
Thelétteri, Female or shee Ferne.
Thélico, a stone like an Oliue-stone, which being beaten to pouder
breaketh the stone in the kidneis.
Thelifóno, the hearbe Libard-bane, the roote whereof doth kill Serpents.
Theligónio, the hearb Culerage, Arse-mart or as some call it the grace of
God.
Thelipteríno, a kind of Fern.
Théma, a theame, an argument, a subiect, a ground or matter supposed
to be written or spoken of. Also a figure set by Astronomers containing
the state of heauen and the planets at any certaine time. Also an
horoscope or consideration of the time of ones birth or natiuity vnder
this or that planet.
Themísia, Oke of Ierusalem or Mille-graine.
Themología, as Ethimología.
Thenáre, the fleshy part of the hand between the thumbe and the middle
finger.
Thendálida, as Sanguinária.
Theodía, Deity, God-head, the word of God, Diuinity. Also Dauids Psalmes.
Theogónia, the Originall, the beginning or generation of the Gods.
Theolẻpto, inspired of God.
Theologáre, to reason of holy things or preach diuinity.
Theología, reasoning of holy things, science of diuinity, profession of holy
things.

Theológico, theologicall, of or pertayning to diuinity or holy things.

Theologizzáre, as Theologáre.
Theólogo, a preacher or professer of diuinity or holy things, a deuine.
Theomachía, fighting against God.
Theómaco, a fighter against God.
Theománte, a professer of Theomantía.
Theomantía, a kinde of diuination or enchanting by abusiue calling vpon
the secret, farfetcht, misterious and wrested names of God.
Theombrotióne, an hearbe.
Theomenia, Gods wrath.
Theonino, a slanderous man.
Theóphilo, a louer of God, louing God.
Theoréma, a theorem, any axiom or vndoubted truth of an arte, namely
that respecteth speculation more then practise.
Theoremísta, a professor of theorems or axioms of vndoubted truth.
Theorẻtico, of or belonging to contemplation. Also as Theórico.
Theoría, contemplation, speculation, deep study, insight or beholding.
Theórica, as Theoría.
Theórica astronomía, that part of stronomy that giueth the view of the
motions of the plants by their Orbes.
Theórico, a speculator or contemplatiue man. Also theoricall rather then
practike. Also belonging to contemplation and view.
Theorítico, as Theorético.
Thephríte, a stone wherein is seene a new mone tipped with hornes.
Therẻo, a birde called a Robin-red-breast.
Theriáca, triacle, a remedy against poson. Also a Viper.
Theriómata, an vlceration or sore about the priuy members, whereby all
the parts about them are corrupted.
Therionárca, an hearbe with a floure like a Rose, whereon if wild beastes
do but tread they become drouzie and sencelesse.
Thẻrme, hot bathes or baines, properly waters that are naturally warme.
Thẻrmíno, a kind of Oyle of Lillies.
Thẻrone, the name of a dogge, as one would say kill-bucke or fierce-looke.
Thẻrpsicóre, one of the nine Muses, as much to say Maiesty.
Thesaurizzáre, as Thesoreggiáre.
Thẻse, a generall question or indefinite argument.
Thesióne, an hearb very bitter in taste vsed in purges. Also a kind of wine
in Egypt.
Thẻsmophórie, certaine feastes vnto Ceres vsed in Athens.
Thesoreggiáre, to lay vp treasure, to heape or hord vp riches.

Thesorería, a treasury, a place where treasure is kept, a secret store-

house.
Thesoriẻre, a treasurer, a treasure-keeper.
Thesóro, treasure, riches, wealth.
Thẻssa, a kind of fish called the Purple-fish.
Thessálico, a kind of purple colour.
Thẻta, the Greeke letter Theta. Also a marke of condemnation to death.
Theúrgia, a kind of diuination or enchanting.
Thía, as Trogéte.
Thiáde, as Ménade, certaine religious women of the order of Bacchus.
Thiéio, the ancient name of Amber.
Thieldóni, a kinde of ambling Iennet bread in Asturia in Spaine.
Thifóne, as Tiphóne.
Thímalo, a kind of dainty fish.
Thímba, the hearbe Winter-sauory.
Thímbra, the hearb Winter-sauory.
Thímbro, a crest in Armory.
Thimẻlea, as Camẻlea.
Thími, a kinde of long and big warts that haue but a small roote.
Thimiamáte, an Altar of Gold in Salomons Temple, on which were burned
sweete perfumes.
Thimióne, as Thími.
Thímo, the hearbe Tyme. Also a kernell or glandule vnder the canell bone,
which in men is slender and in beastes thicke.
Thimóso mẻle, an excellent hony that Bees make of Thime.
Thínnide, the spaning Tunny fish.
Thínno, the Tunny fish, some take it for the Multer.
Thíphe, a kind of corne in Egipt.
Thiphóne, a kind of prodigious storme or tempest comixt with flashes and
lightnings of fire, the firmament seeming to open, some say it is a
blazing star resembling fire, plated or twisted in maner of a wreath very
grimme and hideous to looke on.
Thiriatíno, a kind of faire bluish colour that Painters vse.
Thirsíte, a stone like vnto Corall, which vsed in drinke procureth sleepe.
Thírso, the stalke, the stocke or stem of any hearbe. Also a Iauelin
wrapped about with Iuy, caried about by the Menades or Harlots in the
 feastes and sacrifices of Bacchus. Some haue also vsed the word for the
enciting of a fury or violent motion.
Thisielióne, Water-parsly.
Thispe, an hearbe some take it for Senuie, others for Country-mustard.
Thóe, a kind of wilde beast, some take it for a kind of very great Woolfe.
Thombáre, thump.

Thómbo, a thump.
Thopalóne, a water-rose.
Thoráce, as Toráce.
Thoriphétro, the hearbe Lions-paw.
Thóro, as Tóro, a bull.
Thrácia, a kinde of blacke stone or Agath, which being steeped in water,
burneth, and sprinkled with Oyle, it quencheth.
Thrásia, a wind between the North and the summer-setting of the Sunne.
Thréni di Gieremía, the lamentations of Ieremy.
Threustóne, a kind of sweet gum like Frankincense.
Thriálli, a kind of Primrose or Couslip.
Thridáce, Sper-hawke-hearb or Hawk-weede.
Thridacẻa, the female Mandrágora.
Thríssa, a kind of Purple fish.
Thróno, a throne or seate of maiesty, an Emperours or Kings seate. Vsed
also for a Kingdome.
Thù, Ground-pine or field-cipres. Vsed also for incense or Frankincense.
Thúmbo, a thump.
Thúnnide, as Thínnide.
Thúnno, as Thínno.
Thuríbolo, as Torríbolo.
Tì, to thee, thee. Also from thee. Looke for it in my rules.
Tiáde, as Ptiáde.
Tiáme, any Ouen or Furnace.
Tiáno, any Ouen or Furnace.
Tiára, a turbant or round wreath of linnen as Turks weare on their heads,
but it was properly a round ornament or head-attire that Princes were
wont to weare in Persia.
Tíbia, any kind of flute, pipe or recorder. Also a trumpet. Also the shanke
or shin-bone of a mans legge which, comprehendeth all the space
betweene the knee and the foote, and containeth two bones, of which
the greater is called Tíbia, it hath also beene vsed for a thrashing of
Corne.
Tibiále, a shin-piece, a greaue or leg-harnish. Also a nether-stocke or
knee-hose.
Tibiáre, to sound vpon any flute, a pipe or recorder. Also to thrash corne.
Tibiatóre, a piper vpon pipes, flutes or recorders, a trumpeter. Also a
thrasher.
Tibidrágo, a kinde of strong Spanish wine or sacke, we call it Rubiedauy.
Tibilústri, daies in which men were wont to go with pipes and trumpets, as
it were in procession about their lands.
Tíbis, a birde like a Storke feeding on Serpents egges.
Tibribástio, iargon, gibrish, pedlers french, fusti[an] tongue, rogues
language.
Tíbulo, a kind of Pine-tree.
Tiburíno, a monstruous Sea-fish which hath two rankes of sharpe teeth.

Ticcále, a kind of waight in India.

Ticcióne, hath beene vsed for Tizzóne.
Tichirnábo, a great Titimouse or Oxe-eye.
Tiẻlla, any kind of Tarte-pan or Pudding-pan. Also hold hir or it fast.
Tiẻllo, hold him or it fast.
Tiémo, as Témo, or Temóne.
Tiẻnti buóno, one that can chuse or take hold of what is good.
Tiepidézza, as Tepidézza.
Tiepidíre, as Tepidíre.
Tiẻpido, as Tẻpido.
Tiepidóre, as Tepidézza.
Tiepóre, as Tepidézza.
Tiẻrmíte, a kind of wilde Oliue.
Tífa, an hearbe growing by or in the water.
Tifoláre, as Strilláre.
Tífole, as Stríllo.
Tíglia, as Tília.
Tíglio, ficcólla al gigánte in mẻzo il tíglio.
Tígna, as Tégna. Also as Tárma.
Tignáme, the Lote-tree.
Tignáto, moth-eaten, worme-gnawne.
Tígno, a Mite, a Weeuell or Meale-worme.
Tígnola, as Tígno, but propely a moth that deuoureth cloathes.
Tignosíre, sísco, síto, to become scald-headed.
Tignóso, that hath a scald-head.
Tignosúzzo, as Foramẻllo, a filthy scald-headed or scuruy fellow.
Tignuóla, as Tignóla.
Tigráme, a kinde of sweet gum or perfume.
Tígre, a Tiger, a most swift beast.
Tigríne, a wood whose graine curbeth in and out.
Tigúrio, as Tugúrio.
Tília, the Linden-tree, the teile-tree.
Tílio, as Tília.
Tilóne, as Millepiédi, but properly a vermine called a Cheeselip.
Timállo, as Timálo.
Timálo, a fish called a flower, goodly to looke vpon, and sweet in taste and
smell.
Timariótti, such men as liue vpon Timarros or hold land in Knights-seruice
or Capite, vnder the great Turke.
Timárro, Lands, Farmes, Mannores, Lordships, Liuings or States of new
conquered lands distributed and holden in Capite or Knights seruice
vnder the great Turke by well-deseruing Captaines and souldiers.
Timbáde, witches, hagges or sorceresses which keepe about tombes and
graues, and cut off some members from dead bodies, which they abuse
in their enchantments.
Tímbo, a tombe, a graue, a sepulcher.
Timbrẻo, Watermints.
Timelẻa, a shrub bearing a kind of Pepper.
Timidézza, as Timidità.
Timidíre, dísco, díto, to make or become timorous or fearefull, to possesse
with feare.
Timidità, timidity, feare, timerousnesse, dread, awe, fearefullnesse. Also
basfulnesse.
Tímido, timorous, fearefull, awefull. Also bashfull or faint. Also doubtfull
hearted.
Timiópoli, Higlers, Engrossers or Regraters of Markets, which make victuals
deare.
Timmóte, one that hath his limmes contracted, conuolued or shrunken vp.
Tímo, as Thímo. Also as Tímolo.
Tímolo, a kinde of Water-snake that swimmeth on hir backe, and hath six
teeth.
Timología, as Etimología.
Timonáre, to steere or guide a ship. Also to wagle the taile as some birdes
doe.
Timóne, the sterne, the helm or rudder of a ship. Also the limmer or
beame of a Wagon or Waine or the draught-tree whereon the yoake
hangeth. Also a yoake. Also an axeletree. Also any rafter, beame or
stake laide athwart, among riders Timóne is vsed for the riders hand
that holdes the bridle, for that with it he gouerneth & directeth the
horse.
Timoneggiáre, to steare a ship. Looke Timóne.
Timoniẻre, a Helm-man or Stearer of a ship.
Timonísta, as Timoniẻre.
Timorággine, fearefullnesse, awefullnesse.
Timoráto, indued with feare, respect and awe.
Timóre, feare, dread, fearefullnesse, scare.
Timorosità, fearefulnesse, as Timidità.
Timoróso, timorous, fearefull, awefull.
Timpanía, a bell-pearle.
Timpánico, diseased with a Timpanie.
Timpanísta, a Drumster, a Tabourer.
Timpaníte, a drie timpany or dropsie rising of much windinesse gathered
betweene the skin Peritonéo and the bowels, which smitten vpon with
the palme of the hand maketh a noise like a drumme.
Tímpano, a Timbrell, a Tabour, a drum, a Drum-slade. Also one of the
instruments of hearing. Also a round bottomed dish or drinking mazor.
Also any pearle fashioned like a bell. Also a Printers toole called a
timpine. Also the Water-mill-wheele that in turning about taketh and
deliuereth water. Also a Cog-wheele. Also a kind of bird.

Tímpano a vólta. Looke Campanẻllo.

Tímpano piáno. Looke Campanẻllo.
Timpintáo, a chiefe officer in China as Lieutenant to a Viceroy.
Tímulo, a kinde of Water-snake that hath six feete and swimmeth on hir
backe.
Tína, any kind of tunne or great vat.
Tináccio, as Tína, or as Mastẻllo.
Tinacciuólo, as Mastẻllo, or Tína.
Tinázzo, as Tína, or as Mastẻllo.
Tínca, a fish called a Tench. Vsed also for a fresh water souldier, or
vnexpert Captaine that will haue thirty men with him be it but to digge
vp a Turnep.
Tinchétto, a Crosse or Maine-saile.
Tincóne, as Tencóne.
Tindáridi, the twins Castor and Pollux.
Tinẻllo, a little tub, tray or washing-boule.
Tinẻllánti, ordinary seruing men that neuer wait or eate but in the common
hall.
Tinẻllísta, as Tinẻllánte.
Tinẻllo, any kind of tunne, wine-pipe, hogshead, barrell or tub, a Bruers
vat. Also a hutch, a bing, &c. Also the Common-hall in any noble or
Gentlemans house where ordinary seruants and waiters attend and dine.
Tinésmo, as Tenasmóne.
Tínge, as Ptínge. Also he dieth.
Tíngere, tíngo, tínsi, tínto, to die, to staine or colour into any hew or
colour.
Tínger chi fálla, a kinde of Christmas game where he that misseth is
slurred over the face.
Tiniẻro, hauéua cíglia come árco da tiniẻro.
Tiniuẻlla, a wimble, an augure, a piercer, a drill. Also a neuell, a niuell
among Gunners.
Tiniuẻlláre, to wimble, to pierce, to drill, to boare or niuell thorow.
Tiniuẻlláto, a pice niuelled or newelled or boared. Also boared or wimbled
thorow.
Tinniníre, nísco, níto, to chirpe like a Miskin or Titmouse.
Tinnisóne, a kind of beast with a very rough skinne.

Tinnitíre, tísco, títo, as Tintináre, to sound shrill.

Tinníto, as Tintíno, a shrill sounding.
Tinnúnculo, a Kestrell or Stannell.
Tíno, a kind of wilde Laurell or Bay. Also as Tína.
Tinóccio, a great woodden tunne or Vatte.
Tinóre, as Tenóre.
Tinósa, as Tinẻlla.
Tinózzo, as Tinóccio.
Tínta, a dye, a dying, a hue, a colour, a colouring. Also the water or glasse
of any precious stone.
Tintalhóra, a wanton, wily, tempting lasse.
Tínta néra, blacke dye or inke.
Tintarẻlla, a pretty dye or hue.
Tinténe, a little shrill-sounding bell, a musicall instrument without strings.
Tintilláni, fine cloathes woadded or as we say died in wooll.
Tintinábulo, as Tinténe.
Tintinaménto, as Tintíno.
Tintinánte, tingling, shrill-ringing, iangling, sounding, sharpe or shrill in
sound.
Tintináre, to tingle, to iangle, to gingle, to ring shrill and sharpe as some
bels.
Tintiníre, nísco, níto, as Tintináre.
Tintiníto, as Tintíno.
Tintíno, any kind of shrill tingling, gingling, or sharp sounding, as of
basons or bels.
Tínto, dyed, stained or coloured, dipped or washed in any colour. Also
blushed red.
Tínto in rósso, died in scarlet.
Tínto in grána, died in graine.
Tintóre, a Dyer or Stainer of colours.
Tintoría, a Dye-house, the trade of dying.
Tintúra, any dying, dye or colouring.
Tiórba, a musical instrument that blind men play vpon called a Theorba.
Típha, Typh-wheate or wheate of Rome.
Típha acquática, a flower that flyeth away like the downe of a thistle called
Water-torch, Cats-taile, Red-mace, Dutch-downe, Marsh-beetle or
Marsh-pestle.
Típho, arogancy, haughtinesse of minde.
Tiphóne, as Thiphóne.
Tipicaménte, mistically, shadowingly, figuratiuely-representing.
Típico, misticall, as it were shadowed, and represented by some figure.
Típo, a tipe, a figure, a shadow, a forme, a likenesse or sample of any
thing.
Tipocósmia, a tipe or figure of the world.
Tipografía, the Arte of Printing of bookes.

Tipógrafo, a Printer of Bookes.

Tipúlla, a Water-spider with six feete, that runneth on the water without
sinking.
Tíra, a drawing in length, a pulling.
Tiracóllo, a halter or drawer. Also as Tiradóri.
Tiradóri, the traces that belong to Cart-horses.
Tiralónga, a kinde of boies play in Italie.
Tiráme, all manner of haling or drawing vnto, or shuting and throwing
from one. Looke Tiráre.
Tiraménto, any pulling or hailing. Also any casting or throwing.
Tiranẻllo, a little Tyrant, one that is somewhat tyranous.
Tirannáre, as Tirannizzáre.
Tirannescaménte, tyranously, cruellie, like a tyrant.
Tirannía, tyrany, cruelty, a violent rule for priuate commodity and not the
publike weale.
Tirannicída, a murtherer of a Tyrant.
Tirannicídio, murther of Tyrants.
Tiránnico, tyrannicall, cruell, tyranous.
Tiránnide, as Tirannía.
Tirannizzáre, to tyranize, to be cruell.
Tiránno, in old time it was taken for a King and in good part, but now it is
vsed for a tyrant or cruell Lord and tyranicall King and murtherer. Also a
kind of Eagle or Faulkon called a nine-murtherer.
Tirannóso, full of cruelty and tyranny.
Tiránte, drawing, stretching. Looke Tiráre. Also breeches or hosen in
Pedlers french.
Tiraóro, a Gold-wire-drawer.
Tirapáncia, a great feeder, a stretch-gut.
Tiráre, to draw, to pull, to hale, to plucke, to tug or attract vnto or towards
ones selfe. Also to withdraw or retire. Also to shrinke in. Also to stretch
in or out. Also to throw, to cast, to fling, to hurle, to shoote, to darte, to
sling or pitch from one. Also to yarke, to kicke or winze with ones heeles
as a horse doth. Also to protract, to wire-draw, to prelong, to driue off,
or draw in length. Also to entice, to perswade or draw and bend vnto.
Also to draw or pourtray.
Tiráre a bótta di fíco, to shoote alwaies in one place.
Tiráre ácqua, to draw water.
Tiráre álla strácca, to shoote at the vttermost randam.

Tiráre buón púnto, to cast a good cast at dice.

Tiráre cálci, to kick or winze with ones heeles as horses doe.
Tiráre costiẻro, to shoote wide or sidelin.
Tiráre da cánto, to draw or pull a side.
Tiráre d'árco, to draw or shoote in a bow.
Tiráre d'árme, to play at fence.
Tiráre d'artegliería, to shoot of or with great ordinance.
Tiráre dẻntro délla puntería, as Tiráre dẻntro il uíuo.
Tiráre dẻntro il uíuo, to shoot without the disparte. Look Víuo.
Tiráre di bócca, to beare too hard vpon the bit as some head-strong
horses doe.
Tiráre diccinóue, to cast more then three dice haue, to goe beyond ones
commission.
Tiráre di míra, to shoote leuell, or right.
Tiráre di puntería, that is when a shot is made by the concauity of the
piece, which is to shoote by the disparte, and the disparte is when a
piece of wax or sticke is set vpon the mouth of the piece in an euen line
with the cornish of the breech.
Tiráre drítto, to shoot right, leuell or point-blanke. Also to draw straite.
Tiráre d'vn pẻzzo, the range of a piece. Also to shoote with a piece.
Tiráre fuóra délla puntería, as Tiráre fuóra del uíuo, or di puntería.
Tiráre fuóra del víuo, to shoote without or beyond the disparte.
Tiráre fuóri di puntería, to shoote at randon, which is to make a shot upon
any degree of the quadrant, which doing the piece must be raised above
point blanke.
Tiráre gióia pẻr gióia, to shoote levell, taken by the upper part of the two
cornishes at the breech and mouth of the piece, without helpe of
disparting.
Tiráre inánzi, to draw or shoote forward.
Tiráre indiétro, to draw or shrinke backe.
Tiráre la bárca, to hale or towe a boate.
Tiráre la bárra, to throw or pitch the barre.
Tiráre la gióia, to shoote by the vpper superficies of the cornish of the
mouth of the piece, which the Italians call point blanke, which I take to
be when we shoot with a dispart of both cornishes, that is mouth and
breech.
Tirár l'aiuólo, to allure, to entice, to ensnare, to entrap.
Tiráre l'anciána, to hale, to halse, to tuge, to draw or tow a barge with
cordes.

Tiráre la pága, to receiue or draw pay.

Tiráre la sórte, to cast or draw lots.
Tiráre le cálcie, to pull of ones hosen, id est, to die or yeeld vp the ghost.
Tiráre le póste, to draw stakes.
Tiráre pẻr i críni, to draw to pull or tug by the haires.
Tiráre pẻr il ráso dell'ánima, to shoote vp a iust leuell, which is when the
concauity of the piece both, at mouth and breech is laide vpon a right
line.
Tiráre pẻr il ráso de metálli, to shoote by the extreames of the two
cornishes of mouth and breech, which our Gunners improperly call Point
blanke.
Tiráre pẻr il ráso de metálli in giù, to shoote vnder mettall.
Tiráre pẻr il ráso de metálli in sù, to shoot at randon.
Tiráre pẻr l'ánima del pẻzzo, as Tiráre pẻr il ráso dell'ánima.
Tiráre pẻr linẻa rẻtta, to shoote point blanke or in a direct line.
Tiráre púnto in biánco, to shoot point blanke, that is when a shot is made
by the sight or leuell of both cornishes of the piece, shooting as gunners
tearme it, neither aboue nor vnder mettall, that is vpon no degree.
Tiráre sótto, to draw or shoot vnder.
Tiráre vẻnto, the wind to blow.
Tiráre úna coréggia, to let a girding farte.
Tiráre vn pétto, to let or guird out a fart.
Tiráre vn sásso, to hurle or fling a stone.
Tiráro, as Tiratóre. Also as Carrúccio.
Tiráta, a draught, a pulling or haling vnto. Also a twitch, a plucke, a snap.
Also retiring or withdrawing. Also a shrinking in. Also a stretching out.
Also a throwing, a flinging, a hurleing, a shooting, a darting from one,
&c. Looke Tiráre or Tíro.
Tiráta d'árco, a bow-shoote.
Tiráta di sásso, a stones cast.
Tiratẻlla, any little Tiráta.
Tiratíuo, drawing, pulling. Also attractiue. Also shrinking in or stretching
out. Also casting, hurling or flinging. Looke Tiráre.
Tiráto, drawne, pulled, haled, pluct, tugged or attracted vnto one. Also
retaired or withdrawne, shrunken in or stretched out. Also throwne,
cast, flung, hurled, darted, pitched, or shot from one. Also protracted.
Looke Tiráre.
Tiratóio, a till or drawing box, or any thing that may be drawne. Looke
Tiráre.

Tiratóre, a drawer, a puller, a haler. Also a caster, a hurler, &c. Looke

Tiráre.
Tireliráre, to sing as a Larke.
Tirémo, nè mónti, nè tirémi, nè bóschi, nè sélue.
Tirẻo, a stone of the nature of Scírio.
Tirésio, as much to say, blinde or sightlesse.
Tirimattáre, a play at tables like vnto our english tick-tacke.
Tírio, a kind of purple in graine died at Tirium.
Tíro, a drawing, a pulling, a hailing, a plucking, a tuging, an attracting vnto
or toward ones self. Also a draught, a pul, a pluck, a tug, at witch or
twiching vnto. Also a shrinking in. Also a stretching out. Also a draught
or line in painting or writing. Also a throw, a cast, a fling, a hurle, a
darting, a slinging, a pitching, a shot or shooting or tire from one as a
cast of a stone, a stones cast, a tire of Ordinance, a shoot out of a bow
or tiller. Also a draught at chesse. Also a draught of drinke, a sup of an
egge, a whife of Tabacco. Also a stroke, a bang, a blow, a hit, a nip or
twang with a cudgell. Also a veny at fence. Also a pranke, a feat, a part,
a tricke, a knacke, a deuise, a slight, a conceite, a drift or ayme at a
thing. Also a reach. Also a distance, as for example, a mery pranke, a
nimble feate, and odde part, a tricke of youth, a fine knacke, a cunning
slight, a subtile drift, a rare conceite, a long reach, a farre distance, a
cunning ayme, a tricke at cardes, a cast at dice, a nip on the head. Also
a scoff, a taunt or quip giuen to the quicke, a hitting to the quicke. Also
a dash of a penne. Also a slash or cut with any weapon. Also a yarke of
a whip. Also a kicke or winzing with ones heeles. Also any kinde of
shrinking in or contracting of sinnewes. Looke Tíro di nẻrui. Also he that
first entreth into experience or practise of any feate, arte or science, a
yongue nouice, a beginner or new learner. Also the name of a most
 venemous kinde of Viper, Adder or Serpent, whereof the best Triacle or
Antidotes for poison is made, and thereof taketh it's name.
Tíro alla strácca, a shoote made at randon.
Tirocínio, vnskilfulnesse, lacke of knowledge, experience and wisedome.
Also the first begining, entrance into or exercise of any thing, an
apprentisage.
Tíro di máni, a legerdemaine tricke. Looke Tiráre, and Tíro.

Tíro di nẻrui, any contracting or shrinking of sinnewes. Italian riders say it

is a disease in a horse which commeth of great could and heate,
whereby the great sinnew in the necke is so restrained as the horse
cannot well open his mouth.
Tirománte, a deuiner by cheese.
Tiromantía, a kinde of deuination by cheese.
Tiróne, a yongue nouice, a milke-sop, a faint-hearted souldier, a beginner
to learne.
Tirréno, a kinde of stone which being whole swimmeth, and broken
sinketh.
Tírsi, as Thírso. Also the circle without the wals of a towne, wherein men
may walke for their solace.
Tisána, Ptisan or Barly water sodden.
Tisbína, a kinde of childes play in Italie.
Tisichézza, Ptisiknesse, purcinesse, shortnesse of breath, the cough of the
lungs.
Tisichíre, chísco, chíto, to become purcy, ptisike or grow short winded.
Tísico, ptisike, purcy, short winded.
Tisicúme, as Tisichézza.
Tisicúzzo, somewhat purcy or shortwinded.
Tísigo, as Tísico.
Tistéra, a hood of a cloake or cape of any garment.
Títa, a title in writing, a whit, a iot.
Titalóra, a silly gull, a foolish ninny.
Titáno. Vsed for the Sunne.
Titée, Tickes, Sowes, Woodlice or Cheeslips.
Titilláre, to tickle, to prouoke pleasantlie.
Titillatióne, a tickling, a ticklishnesse.
Titíllico, a tickling. Also ticklish.
Titillóso, ticklish, full of ticklings.
Titimáglio, Milke-thistle, Sea-lectuce or Wolfes-milke, or as some say
Spurge-hearbe.
Titimálo, as Titimáglio.
Titispíssa, a bird called a Wag-taile.
Titoláre, to title, to entitle, to surname, to nickename. Also titular.
Titolatióne, an entitling, a titulation.
Titoláto, titled, entitled, surnamed, nickenamed. Also a man of titles or
dignities.
Títolo, a title, a pricke or point vsed in writing. Also a title as a surname, a
name of dignity. Also a title or inscription of any act or worke.
Titubánte, wauering, staggering. Also stumbling, tripping, or reeling. Also
faltring, stutting or stammering.
Titubánza, a wauering, a staggering, a stumbling, a tripping, a faltring or
stammering.

Titubáre, to wauer, to stagger. Also to stumble or trip. Also to falter, to stut

or stammer.
Titubatióne, as Titubánza.
Tituláre, as Titoláre.
Titulatióne, as Titolatióne.
Título, as Títolo.
Tiuẻrtíno, a kind of building marble.
Tizẻrbétto, as Sorbétto.
Tizzáre, to stir vp the fire or firebrands. Also to irritat, to egge on or
prouoke.
Tízzo, a fire-brand.
Tizzonáre, as Tizzáre.
Tizzóne, any fire-brand, but properly a quenched fire-brand halfe-burned.
Tizzoniéra, a fireforke to stir the fire or fire-brands. Also a busie woman
that stirreth vp quarrels among hir neighbours or other people.
Tlaguáce, a wilde beast that carrieth hir yongue ones in a pouch or bag
vnder hir belly.
Tlási, a rupture or bruising in a mans cods or stones.
Tláspi, the roote or hearbe called Raifort or Country mustard.
Tò, take, hold, catch, take to thee. Also he taketh. Also vsed for two,
thine.
Tóca, a bird in India, of the bignesse of a Rooke with a white breast and
guilt bill.
Tócca, a touch. Also a touching. Also a touch-stone. Also tinzell cloath of
Gold or siluer.
Tócca a tè, it toucheth, concerneth or belongeth to thee.
Tócca a vói, it toucheth or concerneth you, it belongs to you, it fales to
your lot.
Toccadíglio, a kinde of game at tables much like our Tick-tacke.
Tócca d'óro, Gold-tinzell or Tissue.
Toccaménto, any kinde of touching, touch, feeling or stroaking softly.
Toccáre, to touch, to feele. Also to hit, to ioyne close vnto. Also to touch,
to write, to speake or mention any thing by the way. Also to smite, to
strike. Also to guip or tante. Also to appertaine or belong vnto, to fall to
ones lot.
Toccáre ad áltri, to belong, to concerne, to appertaine or fall to others lot.
Toccáre a martẻllo, to ring the bels with a hammer or as we say
backeward.
Toccáre il pólso, to touch or feele ones pulse.
Toccár' in párte, to fall to ones lot or share.
Toccár' in sórte, to fall to ones lot.
Toccár' la máno, to touch, to feele or take by the hand, as we say to shake
hands.

Toccár la vicẻnda, to fall to ones turne.

Toccár la vólta, idem.
Toccáre li tásti, to touch the keyes or frets of an instrument, to touch to
the quicke.
Toccáre sul víuo, to touch to the quick, to make one smart.
Toccáta, a touching, a feeling.
Toccáta d'vn músico, a preludium that cunning musitions vse to play as it
were voluntary before any set lesson.
Toccáto d'óro, as Tócca d'óro.
Tocchétti, a kind of meat sliced, shread or cut in steakes or collops. Also a
kind of little waight in Italy.
Tócco, with a close o, a touch, a feeling. Also a veny at fence, a hit, a
stroake or knocke. Also a stroke of a bell or clocke. Also a touch-stone.
Also a nod, a becke, a signe or glance at any thing. Also a certaine
childes play vsed in Italie. Also touched or felt. Also hit or stroken. Also
belonged or fallen to ones lot.
Tócco, with an open o, any piece, scrap, luncheon, cob, collop,
mammocke, cut or shiue namely, of bread and cheese.
Tócco di campána, a knocke, a stroke, a knell, a peale or toule vpon a bell.
Toculáre, to touch, or feele.
Tódo, a Muskin or Titmouse.
Tófo, a kind of soft, sandy or mouldering stone.
Tofóso, sandy, mouldering.
Tóga, a goune, a roabe or long garment which the Romanes did euer vse
to weare in times of peace.
Togále vẻste, a robe-like garment.
Togáto, gouned, that weareth a goune.
Tóga viríle. Looke viríle.
Tógliere, tólgo, tólsi, tólto, to take, to receiue, to accept. Also to take
away, to remooue from. Also to free from.
Tógliere a fítto, to take to farme or rent.
Tógliere a nuólo, to hire as a ship or horse.
Tógliere il vánto, to depriue another of vant or glory. Also to vant or boast
of.
Tógliere in iscámbio, to take in exchange. Also to mistake one for another.
Tógliere l'assúnto, to take the charge or burthen vpon himselfe, to
assume.
Tógliere vía, to take away or from, to depriue of, to remooue away.
Tógna, a fishing rod, an angling rod. Also a slut, a flurt, a driggle-draggle.
Tóla, any kind of flat boord, table or planke. Also a broad flat stone, any
thin plate.
Tólda, the fights about a ship. Some take it for the vpper decke.
Tóle, little thin boordes. Also as Tólle.
Toleóne, an engine to draw vp water as Bruers and Dyers vse with a great
waight at the end, it is called a sweepe.
Tolerábile, tolerable, that may be suffered, endured or borne.
Tolerabilità, tolerablenesse.
Tolerándo, as Tolerábile.
Toleránza, toleration, sufferance, enduring, patience. Also permission.
Toleráre, to tolerate, to suffer, to endure, to beare with. Also to permit.
Also to nourish or maintaine himselfe.
Toleratióne, as Toleránza.
Tólga Dío, God forbid.
Tólla, thin plate of Brasse or Latten, as Tóla.
Tólle, certaine glanduls, kernels, or swellings in the iawes. Also as Tólla.
Tollenóne, as Toleóne.
Tolleóne, as Toleóne.
Tóllere, as Tógliere.
Tolléta, theft, larceny, pilfering, filching, purloyning. Also any booty.
Tolléto, as Tolléta.
Tólto, taken, receiued, accepted. Also taken from or away. Looke Tógliere.
Tólto súso, taken vp.
Tolutiloquẻnza, a kind of swift or hasty speaking.
Tomacẻlla, a kind of Hagasse or Hogges pudding. Also a kind of meate
made of the plucke, giblets, purtenances or tripes of beastes. Also that
meat which we call Oliues of rosted Veale.
Tomáia, the rand of a shooe. Also a tack of clouting leather. Some take it
for the vpper leather.
Tománo, a coine in Persia worth about twenty crownes.
Tomára, as Tomáia.
Tomáre, to tumble, to rumble, to roule or fall downe groueling.
Tomasẻlla, as Tomacẻlla.
Tómba, a tombe, a graue, a sepulcher, a monument or hollow darke place.
Tombáre, to entombe, to bury, to enter or laie in a graue.
Tomboláre, as Tomáre, to tumble.
Tomboláta, a tumbling tricke.
Tómbolo, a tumbling tricke. Also a measure of corne about a bushell of
ours.
Tomẻnto, flockes to fill Matarasses.
Tómici dẻnti, the fore-teeth.
Tomíno, a kind of cheese.
Tómo, a downefal, a tumbling, a rumbling, a rowling or falling downe.
Tómo, a tome or volume of a booke, where diuers workes of one Author
be deuided into sundry parts.
Tomoláre, as Tomáre, to tumble downe.
Tómolo, as Tómbolo. Also as Tómo.
Tomothuriána, the sword-fish, or Emperour of the Sea.
Tónaca, any coate or iacket, a sleeuelesse coate. Also a womans peticote
or kirtle and vpper sauegard. Also any couer like a pill, a rind or vpper
skin. Also a coate that couereth the eie, whereof there be foure sortes,
the first Cornẻa, which is white and resembleth horne, the second Vuẻa,
which is like a Grape-kernell, the third Vitrẻa, which resembleth glasse,
the fourth Cristallína, which in clearnesse resembleth Christal. Also a
kind of Gillie-floure.
Tonánte, thundering. Also roaring, rumbling, ratling, loud-sounding, far-
sounding, shrill. Also vsed for God.
Tonáre, to thunder, to rattle, to rumble, to crackle, to rore, to make a
great noise or terrible sound. Also to tune any instrument or voice. Also
to accent or set to any tune.
Tonarióne, a kind of shalme that hath a very sweet and shrill sound.
Tónda, a round. Also as Rónda.
Tondáre, to round, to make round. Also to circle or compasse. Also as
Tóndere.
Tondarẻllo, pritty round, circular.
Tondeggiáre, to round or make round.
Tóndere, tóndo, tondéi, tondúto or tóso, to pare or make round, to pole,
to not, or cut ones haire, to sheare sheepe, to shaue a frier, to clip
money, to crop corne, to reape Barly, to mow Hay, to lop trees, to top
flowres, to brouze or nibble round about. Tóndere monéta, to clip
money, to pare ones nailes.
Tondézza, roundnesse, rotundity.
Tondíno, as Fondẻllo, a funnell. Also a little round or fruite-trencher.
Tondità, as Tondézza.
Tonditóre, a shauer, a notter, a poler, a clipper, a barber. Looke Tóndere.
Tonditúra, a shearing, a notting, a poling, a rounding, a clipping. Looke
Tóndere.
Tóndo, a round, a circle, a compasse. Also a round trencher, a round plate
or little dish. Also a mans bum, trill or arse. Also a Lump-fish. Also a
round wheeling blow. Also rounded or made round. Also a shallow brain
or a logger-head. Also il tondo, hath beene vsed for the round world,
the whole vniuerse.

Tóndo délla lúna, the full of the Moone.

Tóndo sénza pélo, a yongue wenches quaint. Also a yongue mans bum.
Tondúto, pared, made round, poled, notted, shorne, shauen, clipped,
cropt, reaped, mowed, lopt, topt, brouzed. Looke Tóndere.
Tónega, as Tónica, as Tónaca.
Toneggiánte, as Tonánte.
Toneggiáre, to thunder, to rattle, to rumble.
Tonẻllo, any kinde of tunne or barrell. Also a Lump-fish.
Tónia, a Tunny-fish.
Tónica, as Tónaca.
Tonicáre, to goune or put on a Tónaca.
Tonicẻlla, a body or waste-coate, any little Tónaca.
Tonína, a meate made of Tunny fish.
Tonitrábile, that may thunder or crackle.
Tonitrále, wherein is any thundering noise.
Tonitránte, as Tonánte.
Tonitráre, to thunder, as Tonáre.
Tónitro, any thunder or crackling noise.
Tónne, for Tógline, take some of it.
Tonnẻlláta, a tunne weight or burthen.
Tonnéra, the fishing of Tunny.
Tonnína, Tunny fish salted or dressed to be eaten.
Tónno, a Tunny fish.
Tóno, the space betweene the earth and the moon, so called of
Pythagoras. Also thunder or any rumbling and crackling noise. Also any
 tune, note, ayre, accent or sound.
Tonollétto, a wooden cace in forme of a barrell, wherein Gunners put haile
shot to shoot in morter pieces.
Tonsíle, that may be shorne, notted, rounded, clipped or cropped. Looke
Tóndere.
Tonsílle, spungeous kernels in mens throates.
Tonsóre, as Tonditóre.
Tonsúra, as Tonditúra.
Toos, a Serpent which sometimes is hairy and sometimes not.
Tópa, I take or hold the bye when one plaies at dice.
Topáio, a Rat-catcher. Also a Mouse-trap.
Topárca, a Gouernor of any Shire or Prouince vnder another, as Sátrapo.

Toparchía, a Territory of a Shire or gouernment of a Prouince.

Topáre, as Toppáre.
Topátio, a yellow stone called a Topace, shining very cleare in the darke.
Topázo, as Topátio.
Topiária, the tame or garden Brank-vrsine. Also the arte or feate of making
images. And also Arbours of trees, branches or flowers.
Topiário, a garland, a traile or harbour of floures for garnishment.
Tópica, a part of Logike noting the places of inuention. The arte of
inuention, or finding out of arguments. Also pertaining to the places of
inuention or arguments. Also topikes or bookes that treate or speake of
places of inuention, in and touching Logike.
Topicẻllo, any little Mouse, or Rat.
Topinára, a Mole-hill. Also a Rats-nest.
Topinári, the scratches in horses pasterns.
Topíno, a little Mouse or Rat.
Tópo, a Mouse, a Rat. Also a Sea-tortoise.
Tópo aláto, a night-bat, a keare-mouse.
Topógrapho, a describer of places.
Topographía, a description of places.
Topolíno, a little Mouse or Rat.
Topomaríno, a Whale-guide or Sea-tortoise.
Toporágno, a Night-bat. Also the hardie-shrew.
Toporẻllo, a little Mouse or rat.
Tóppa, the spring or ward of a spring-locke. Also a latch or wodden locke
of a poore Country house. Also a dripping pan. Also vsed for Tópa.
Toppáre, to counter-shocke or giue a counter-buffe. Also to finde or meete
withall by chance. Also to snatch or take away. Also to set, to cast at, to
plaie at or hold the by or vie at any game namely at dice. Also to put to
a dore and make it fast with a haspe or latch or wodden locke.
Tóppo, a counterbuffe, a counter shocke at Tilt.
Tór, as Tógliere, or Tórre.
Toráce, the breast or bulke of a man. Also the middle-space betweene the
neck and the thighs, also a placket, a stomacher, or brest plate for the
body.
Tór' a crẻdito, to take vpon credite.
Torághe, the hollow of the eyes, the eye-holes.
Toragóntia, as Taragóne.
Tór bándo, to take banishment.
Torbánte, as Turbánte.
Tór báscio, to take a kisse.
Torbidáre, to trouble, to disturbe, to confound, to pudle or make water
thicke.

Torbidézza, disturbance, troublednesse, thicknesse or muddinesse of

water.
Torbidíre, dísco, díto, as Torbidáre.
Tórbido, troubled, thick, not cleare, muddie, also foule, duskie, gloomie,
obscure, also turbulant or tempestuous.
Torbinále, fashioned brode aboue, and small below like a top or a gigge.
Torbináre, as Turbináre.
Tórbine, as Túrbine.
Torbinẻo, as Turbinẻo.
Tórbo, troubled, thicke. Also as Túrbine.
Torboláre, as Torbidáre.
Torbolẻnte, turbulent, stiring, vnquiet, stormy, blustering, troublous,
contentious.
Torbolẻnza, turbulence, disturbance, bluster, vnquietnesse, confusion.
Torcalétto, a Turners toole that he turnes and workes with. Also a little
turne.
Tórce, a torche, a taper or linke.
Torcéllo, a wreath or wisp of any thing that women set vpon their heads to
carrie vpon.
Tórcere, tórco, torcéi, torciúto, or tórto, to wrest, to wreath, to bend, to
bow, to retorte, to make crooked. Also to twist, to wrap, to winde in or
wherle about. Also to wring, to presse or squeese out, also to crisp, to
curle or frizle. Also to torture or torment.
Tórcere il grífo a qualcúno, to make a scornefull mouth, or turne ones face
from one.
Torcéuole, wrestable, to be wrested. Looke Tórcere, as Torcíbile.
Tórchia, a torch or linke. Also a kind of measure of liquid things.
Torchiáre, to presse in a winding presse. Also to turne or wrest a presse.
Torchiẻlláre, as Torchiáre.
Torchiẻllo, any litle winding presse.
Torchiéro, a torchmaker, or bearer.
Torchiétto, any little torch or linke or taper. Also any little winding presse.
Torchíno, an azure, turkie or blue colour.
Tórchio, a torch or linke. Also any kind of winding presse, a wine or
printers presse.
Tórcia, any kinde of torch or linke.
Torciáre, as Torchiáre, to winde about.
Torciáro, a Torch-bearer or Maker.
Torcíbile, that may be turned, bended or wrested, that is pliable.
Torcicóllo, a Hickway or Wrienecke. Vsed for a dissembling or puling
Puritane.
Torciétta, any little Torch or Taperlight.

Torcifécci, a Grape or Crab-presse.

Torcifécciole, the dregges of wrung Grapes.
Torcigliáre, as Attorcigliáre.
Torcigliatúra, as Attorcigliatúra.
Torcimánno, an interpreter, a trouch-man, a spokes man. Also a good lad
or good fellow.
Torcimáno, as Torcimánno.
Torciménto, as Torcitúra, a wresting.
Torcíno, any little winding presse as for letters or napkins. Also a little
torch.
Torcitóio, the pin, wrest, bar, handle or winder of any winding presse.
Torcitóre, a wrester. Looke Tórcere.
Torcitúra, a wresting. Looke Tórcere.
Torciúto, wrested, bended, made crooked, corbe, retorted. Looke Tórcere.
Tórco, as Tórcolo. Also vsed for a chaine or a wreath.
Torcoláre, to presse as wine is.
Tórcolo, a wine or Oyle presse. Vsed also for the braine tunnell.
Tór comiáto, to take leaue.
Tór credẻnza, to take assay.
Torculáre, as Torcoláre.
Tór d'a dósso, to take from ones backe.
Tordáio, a place to keepe Blacke-birdes in.
Tór dál móndo, to take from the world.
Tordarẻllo, any little Tórdo.
Tordẻila, a kind of water-foule.
Tór délle máni, to take from out ones hands.
Tordiglióne, a kind of dance in Spaine.
Tordilióne, the seede of Seseli of Candia.
Tór di mẻzzo, to take in the middle.
Tór di míra, to hit leuell.
Tór di víta, to take from life.
Tórdo, a Thrush, a Thrussell, a Mauis, an Owsell or Blackebird. Also a fish
like a Tench, of a russet colour, with purple and white spots on his belly
called a Sea-thrush.
Toreatóre, a bull-baiter.
Torẻllo, a yongue Bull, a Runt, Yearling or Steere. Also a Farriers
instrument to cleanse and feele horses sores.
Toreutíce, the arte of grauery or embossing.
Toríbolo, a censer to burne incence in.
Torítio, bullish or rammish.
Tór la víta, to take ones life.
Tór la vólta, to take ones turne.
Tór licẻnza, to take leaue.
Tórlo, a top, a gigge, a nunne or whirlegig that children play with. Also a
spinners werule to put on a light spindle. Also the yolke of an egge. Also
a twirle or round turning tricke in dancing.

Torlorù, a ninny or giddy headed foole.

Tórma, as Túrma.
Tormentáre, to torment, to torture.
Tormentatóre, a tormenter, a racke-master.
Tormentéuole, tormentable.
Tormentílla, Tormentill, Septfoile or Seauen-leaued grasse.
Tormentína, as Trementína.
Torménto, torment, torture, affliction, smarting paine, any racke or engine
to torment withall. Also any kinde of battering engine of war. Also a
wrining or griping of the belly.
Tormentóso, grieuous or full of torments.
Tormína, a wringing or fretting in the guts.
Torminále, that doth or may cause wringing in the guts.
Torminóso, that wringeth or fretteth the guts, one that hath torment in his
guts.
Tór móglie, to take a wife.
Tormóra, as Túrma.
Tornadúra, a kind of measure in Italy.
Tornáio, a turner. Also a kinde of birde.
Tornalẻtto, the valences of a bed, but namely such ornaments as in Italy
are vsed round about the lower part of bedsteads.
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com