Professional Documents
Culture Documents
PDF Medical Image Processing Reconstruction and Analysis Concepts and Methods Second Edition Jiri Jan Author Ebook Full Chapter
PDF Medical Image Processing Reconstruction and Analysis Concepts and Methods Second Edition Jiri Jan Author Ebook Full Chapter
https://textbookfull.com/product/biota-grow-2c-gather-2c-cook-
loucas/
https://textbookfull.com/product/guide-to-medical-image-analysis-
methods-and-algorithms-2nd-edition-klaus-d-toennies-auth/
https://textbookfull.com/product/computer-vision-and-image-
analysis-digital-image-processing-and-analysis-4th-edition-scott-
e-umbaugh/
https://textbookfull.com/product/animal-behavior-concepts-
methods-and-applications-second-edition-nordell/
Introduction to Medical Image Analysis Rasmus R.
Paulsen
https://textbookfull.com/product/introduction-to-medical-image-
analysis-rasmus-r-paulsen/
https://textbookfull.com/product/digital-image-enhancement-
restoration-and-compression-digital-image-processing-and-
analysis-for-true-epub-scott-e-umbaugh/
https://textbookfull.com/product/image-processing-and-gis-for-
remote-sensing-techniques-and-applications-second-edition-liu/
https://textbookfull.com/product/image-processing-dealing-with-
texture-second-edition-maria-petrou/
https://textbookfull.com/product/methods-and-techniques-for-fire-
detection-signal-image-and-video-processing-perspectives-1st-
edition-cetin/
Medical Image Processing,
Reconstruction and Analysis
Taylor a nd F ra ncis Seri es in S ignal Pro cessin g an d Co mmu ni cation s
Series Editor
Fathi E. Abd El-Samie
Menoufia University, Egypt
Speech Processing
A Dynamic and Optimization-Oriented Approach
Li Deng, Douglas O’Shaughnessy
Jiří Jan
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2020 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works
Printed on acid-free paper
International Standard Book Number-13: 978-1-138-31028-5 (Hardback)
This book contains information obtained from authentic and highly regarded sources. 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 acknowledged,
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
photocopying, 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, please access www.copyright.com
(www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of
users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment
has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.
v
vi Contents
Index.................................................................................................................................... 557
Preface
Beginning with modest initial attempts in, roughly, the 1960s, digital image processing has
become a recognized field of science, as well as a broadly accepted methodology, to solve
practical problems in many different kinds of human activities. The applications encompass
an enormous range, starting perhaps with astronomy, geology, and physics, via medical, bio-
logical, and ecological imaging and technological exploitation, up to the initially unexpected
use in humane sciences (e.g., archaeology or art history). The results obtained in the area of
digital image acquisition, synthesis, processing, and analysis are impressive, though it is
often not generally known that digital methods have been applied. The basic concepts and
theory are, of course, common to the spectrum of applications, but some aspects are more
emphasized and some less in each particular application field. This book, besides introdu-
cing general principles and methods, concentrates on applications in the field of medical
imaging, which is specific for at least two features: biomedical imaging often concerns
internal structures of living organisms inaccessible to standard imaging methods, and the
resulting images are observed, evaluated, and classified mostly by non-technically oriented
staff.
The first feature means that rather specific imaging methods had to be developed,
namely, tomographic modalities that are entirely dependent on digital processing of meas-
ured pre-image data and utilize rather sophisticated theoretical backgrounds stemming from
the advanced signal theory. Therefore, development of new or innovated image processing
approaches, as well as the interpretation of more complicated or unexpected results, requires
a deep understanding of the underlying theory and methods.
Excellent theoretical books on general image processing methods are available, some of
them mentioned in the references of this book. In the area of medical imaging, many books
oriented toward individual clinical branches have been published, mostly with medically
interpreted case studies. Technical publications on modality-oriented specialized methods
are frequent, either original journal papers and conference proceedings or edited books,
contributed to by numerous specialized authors summarizing recent contributions to
a particular field of medical image processing. However, there may be a niche for books
that respect the particularities of biomedical orientation while providing a consistent, theor-
etically reasonably exact, and yet comprehensible explanation of the underlying theoretical
concepts and principles of methods of image processing as applied in the broad medical field
and other application fields.
This book is intended as an attempt in this direction. It is the author’s persuasion that
a good understanding of concepts and principles forms a necessary basis to any valid meth-
odology and solid application. It is relatively easy to continue studying and even designing
specialized advanced approaches with such a background; on the other hand, it is extremely
difficult to grasp a sophisticated method without understanding the underlying concepts.
Investigating a well-defined theory from the background makes the study enjoyable; this
aspect was part of the foundation of the concept for the book.
This book is primarily for a technically oriented audience (e.g., staff members in the med-
ical environment, interdisciplinary experts of different (not necessarily only biomedical)
orientations, and graduate and postgraduate engineering students. The purpose of the book
is to provide insight; this determines the way the material is treated: the rigorous mathemat-
ical treatment—definition, lemma, proof—has been abandoned in favor of continuous
explanation, in which most results and conclusions are consistently derived, though the der-
ivation is contained (and sometimes perhaps even hidden) in the text. The aim is for the
xvii
xviii Preface
reader to become familiar with the explained concepts and principles and acquire the idea
of not only believing the conclusions, but also checking and interpreting every result him-
or herself, though perhaps with informal reasoning. It is also important that all the results
be interpreted in terms of their “physical” meaning. This does not mean that they be related
to a concrete physical parameter, but rather that they are reasonably interpreted with the
purpose of the applied processing in mind (e.g., in terms of information or spectral con-
tent). The selection of the material in the book was based on the idea of including the
established background without becoming mathematically or theoretically superficial, while
possibly eliminating unnecessary details or overly specialized information that, moreover,
may have a time-limited validity.
Though the book was primarily conceived with the engineering community of readers in
mind, it should be readable also to technically inclined biomedical experts. It is, of course,
possible to successfully exploit image-processing methods in clinical practice or scientific
research without becoming involved in the processing principles. The implementation of
imaging modalities must be adapted to this standard situation through an environment in
which the nontechnical expert would not feel image processing to be a strange or hostile
element. However, the interpretation of the image results, in more complicated cases, as well
as the indication of suitable image-processing procedures, under more complex circum-
stances, may be supported by the user’s understanding of the processing concepts. It is
therefore a side ambition of this book to be comprehensible enough to enable appreciation
of the principles, perhaps without derivations, even by a differently oriented expert.
It should also be stated what the book is not intended to be. It does not discuss the med-
ical interpretation of the image results; no casuistic analysis is included. Concerning the
technical contents, it is not a theoretical in-depth monograph on a highly specialized theme,
which would not be understandable to a technically or mathematically educated user of the
imaging methods or a similarly oriented graduate student. Such specialized publications
may be found among the references. Finally, while the book may be helpful even as a daily
reference to concepts and methods, it is not a manual on application details and does not
refer to any particular program, system, or implementation.
The content of the book has been divided into three parts. The first part, “Images as
Multidimensional Signals,” provides the introductory chapters on the basic image-
processing theory. The second part, “Imaging Systems as Data Sources,” is intended as an
alternative view on the imaging modalities. While the physical principles are limited to the
extent necessary to explain the imaging properties, the emphasis is on analyzing internal
signals and (pre)image data that are to be consequently processed. With respect to this
goal, the technological solutions and details of the imaging systems are also omitted. The
third part, “Image Processing and Analysis,” starts with tomographic image reconstruction,
which is of fundamental importance in medical imaging. Another topical theme of medical
imaging is image fusion, including multimodal image registration. Methods of image
enhancement and restoration are treated in individual chapters. The next chapter is devoted
to image analysis, including segmentation, as a preparation for diagnostics. The originally
concluding chapter, on the image-processing environment, briefly comments on hardware
and software exploited in medical imaging and on processing aspects of image archiving
and communication, including principles of image data compression.
With respect to the broad spectrum of potential readers, the book was designed to be as
self-contained as possible. Although a background in signal theory would be advantageous,
it is not necessary, as basic terms are briefly explained where needed. Each part of the book
has a list of references, containing literature used as sources or recommended for further
study. The citation of numerous original works, although their influence on and
Preface xix
contribution to the medical imaging field are highly appreciated, was mostly avoided, unless
they served as immediate sources or examples.
The author hopes that (in spite of some ever-present oversights and omissions) the reader
will find the book’s content to be consistent and interesting, and studying it intellectually
rewarding. If the basic knowledge contained within becomes a key to solving practical
application problems and to informed interpretation of results, or a starting point to inves-
tigating more advanced approaches and methods, the book’s intentions will have been
fulfilled.
Jiří Jan
Brno, Czech Republic
xxi
Author Biography
Jiří Jan is a full professor of signal and image processing within the Department of Biomed-
ical Engineering at Brno University of Technology. As the Head of the Department (1990 –
2010) he constituted novel concepts of MSc and PhD study in Biomedical and Ecological
Engineering, and in Biocybernetics. During his professional life, he served here and abroad
(The Netherlands, Denmark, United Kingdom, Belgium, Estonia, etc.) as a university
teacher and researcher.
Since 1980-th he concentrated his scientific interest in the area of biomedical image ana-
lysis. His publication record lists over 280 research papers in scientific journals and confer-
ence proceedings besides several books on digital signal and image processing (VUTIUM,
IEE Press, CRC Taylor&Francis). His activities encompass memberships in editorial boards
of international journals (Springer, IEEE, EURASIP, Hindawi) and in conference boards
(IEEE-EMBC, IEEE-ICASSP, IEEE-ISBI, EURASIP-EUSIPCO, IFMBE, WOC and
others); he was also organizing and chairing the series of 20 biennial international confer-
ences BIOSIGNAL (1978-2012) under the patronage of IEEE-EMBS and EURASIP asso-
ciations. Several teams of his doctoral students were working in the image processing area
under his guidance, namely the Brno section of the national research centre DAR (2005-
2011) and presently the team cooperating with the firm Philips HealthCare Nederland since
2008. His present research interests in medical image data processing concern namely multi-
modal image fusion and diagnostic feature analysis.
xxiii
Part I
Images as Multidimensional
Signals
Part I provides a theoretical background for the book. It introduces the concept of still
images interpreted as two-dimensional signals, as well as the generalization to a multidimen-
sional interpretation of moving images and three-dimensional (spatial) image information.
Once this general notion is introduced, signal theoretical concepts, after generalization to
the two-dimensional or multidimensional case, can be utilized for image processing and
analysis. This concept proved very successful in enabling the formalization (and conse-
quently optimization) of many approaches to image acquisition, processing, and analysis
that were originally designed as heuristic or even thought not to be feasible.
A characteristic example comes from the area of medical tomographic imaging: the intui-
tively suggested heuristic algorithm of image reconstruction from projections by back projec-
tion turned out to be very unsatisfactory, giving only a crude approximation of the proper
image, with very disturbing artifacts. Later, a relatively complex theory (see Chapter 9) that
led to a formally derived algorithm of filtered back-projection was developed. This algorithm,
widely used nowadays, is theoretically correct and provides very good images, even under
practical limitations. Both algorithms are quite similar, the only difference being the filtering
of each individual projection added to the original procedure in the later method—seemingly
an elementary step, but probably impossible to discover without the involved theory. The
alternative methods of image reconstruction from projections rely heavily on other aspects of
the multidimensional signal theory as well.
Part I introduces the basic image processing concepts and terminology needed to under-
stand further sections of the book. Broader and deeper treatment of the theory can be
found in the numerous literature that is partly listed in the references to this section (e.g., in
[4], [5], [6], [18], [22], [23], [25], [26]. Other sources used but not cited elsewhere are [1], [2],
[8], [12], [14]–[17], [19], [21], [24]).
In the context of the theoretical principles, we shall introduce the concepts of two-
dimensional systems and operators, two-dimensional transforms, and two-dimensional
stochastic fields. The text is meant to be self-contained: the necessary concepts of one-
dimensional signal theory will be briefly included but without detailed derivations.
A prior knowledge of signal theory elements, though definitely advantageous, is thus not
1
2 Medical Image Processing, Reconstruction and Analysis
necessary. With respect to the purpose of the book, we shall mostly limit ourselves to the
two-dimensional case; generalization to three- and four-dimensional cases is rather
straightforward and will be mentioned where necessary.
This theoretical section is subdivided into two similarly structured chapters. The first
chapter deals with the theory of images in continuous space, often briefly denoted as ana-
logue images. Besides being necessary as such, because some later derivations will need
this concept, the theory seems also to be more easily comprehensible, thanks to intuitive
interpretations—humans perceive the analogue image. The second chapter deals with dis-
crete images (i.e., discrete-space (sampled) images, the values of which are also quantized)
as only these can be represented in and processed by computers. The reader should realize
that, on one hand, there are many similarities and analogies between the worlds of continu-
ous and discrete images; on the other hand, discretization changes some basic theoretical
properties of the signals and operators. Therefore, the notion of a “densely enough” sampled
discrete image being equivalent to its analogue counterpart is false in principle and often mis-
leading, although the similarities between the areas may occasionally be utilized with advan-
tage; it includes the fact that all real-world images can be represented digitally without any
loss in information content.
1 Analogue
(Continuous-Space) Image
Representation
The physical meaning of the function value and variables is arbitrary. In the present con-
text, such function is usually interpreted as spatially variable brightness (or gray-scale
degree) of a planar still image, dependent on the position determined by the two coordin-
ates x, y forming the position vector r = (x, y). With respect to the interpretation of vari-
ables, the function is often denoted as a spatial-domain function. Theoretically, it is
possible to consider spatially unlimited images, covering the complete infinite plane (x,y);
practically, however, image size (definition extent) is always limited, usually to a rectangle
with the origin of coordinates in the center, for example,
x 2 hxmax ; xmax i; y 2 hymax ; ymax i: ð1:2Þ
A function of the type in Equation 1.1 may be represented in different ways, as can be seen
in Figure 1.1, either as a gray-scale image or as a surface, the z-component of which is
given by the function value, or by means of profiles—one-dimensional functions describing
the function values on a curve f (x,y) = 0, often a line ax + by + c = 0, commonly horizon-
tal or vertical (e.g., for one of the spatial variables x or y constant).
We shall deal mostly with gray-scale images, as they form the output of practically all
medical imaging modalities, so that the above definition should suffice. In medicine, color
is mostly used only for emphasizing the contrast of originally gray-scale images via false
colors or for labeling, neither usually using formally derived algorithms to determine the
colors. Nevertheless, it should be mentioned that a color image can be represented by
a vector function
the components of which describe the brightness of the individual color components, e.g.,
red, green, and blue (differently colored or interpreted triads of components are also used,
e.g., brightness, hue, and color saturation). Each component constitutes a gray-scale image,
and it may be treated as one. In real-color image processing, although still rarely appearing
in medical applications, the images may be treated as vector valued, thus the correlation
among the color components utilized.
3
4 Medical Image Processing, Reconstruction and Analysis
0
0
200
400
400 500
200 300
0 100
0 100 0
0 200 400 0 200 400 0 200 400
FIGURE 1.1 Different representations of a two-dimensional signal: gray-scale image, surface represen-
tation, and profiles along indicated horizontal, vertical, and diagonal lines.
Although natural images are real valued, the generalization to the concept of complex-
valued images is useful theoretically1 and may even simplify computations in some cases.
Again, each part (real and imaginary) can be interpreted as a gray-scale image.
The two-dimensional concept (Equation 1.1) may be generalized to the case of
a multidimensional signal
f ðxÞ; ð1:5Þ
1
Computed images may become complex valued (e.g., in magnetic resonance imaging (MRI)).
Another random document with
no related content on Scribd:
— Sepä vasta aasi! — huudahti Ivan ruveten hermostuneesti
nauraa hohottamaan ja ikäänkuin voimiaan ponnistellen yhä jotakin
harkiten. — Eikö ole samantekevää, makaako ikuisesti vai kulkeeko
kvadriljoonan virstaa? Siinähän on kulkemista biljoonaksi vuodeksi?
— Heti kun hänelle avattiin paratiisin ovet ja hän astui sisälle, niin
hän, ennenkuin oli ollut siellä kahta sekuntiakaan, — ja tämä kellon
mukaan sanottuna, kellon mukaan (vaikka minun käsitykseni
mukaan hänen kellonsa oli täytynyt jo kauan sitten hajota
alkutekijöihinsä hänen taskussaan matkalla), — ennenkuin oli ollut
siellä kahta sekuntiakaan, hän huudahti, että noitten kahden
sekunnin takia kannatti kulkea ei vain kvadriljoona, vaan kvadriljoona
kvadriljoonaa, vieläpä se määrä kvadriljoonannessa potenssissa!
Sanalla sanoen hän lauloi »hoosiannan» ja liioitteli vahvasti, niin että
toiset siellä, jotka olivat vähän jalommin ajattelevia, eivät ensiaikoina
tahtoneet edes antaa kättä hänelle: liian rajusti hän oli siirtynyt
vanhoillisten joukkoon. Venäläinen luonne. Toistan: tämä on
legenda. Mistä hinnasta osti, siitä möikin. Sellaisetpa käsitykset siis
meillä siellä vallitsevat kaikista näistä asioista.
— Ole vaiti, petturi, minä tiesin ennen kuin sinä, että se on Aljoša,
minä aavistin hänen tulonsa, eikä hän tietystikään suotta tule,
tietenkin hänellä on »uutinen!»… — huudahti Ivan raivostuen.
Kolkutus jatkui. Ivan tahtoi rientää ikkunan luo; mutta oli kuin jokin
olisi äkkiä sitonut hänen jalkansa. Hän ponnisti kaikki voimansa
katkaistakseen siteensä, mutta turhaan. Kolkutus akkunaan kävi yhä
voimakkaammaksi ja kuuluvammaksi. Viimein siteet yhtäkkiä
katkesivat ja Ivan Fjodorovitš hypähti sohvalta. Hän katseli hurjasti
ympärilleen. Molemmat kynttilät olivat miltei loppuun palaneet,
juomalasi, jonka hän juuri oli viskannut vierastansa vastaan, seisoi
hänen edessään pöydällä, eikä vastapäisellä sohvalla ollut ketään.
Koputus ikkunan kehykseen jatkui tosin itsepintaisesti, mutta ei
ollenkaan niin kovasti kuin hänestä äsken unessa oli tuntunut, vaan
päinvastoin hyvin hillitysti.
— Tämä ei ole unta! Ei, minä vannon, ettei se ollut unta, se kaikki
oli äsken todella täällä! — huudahti Ivan Fjodorovitš, syöksyi ikkunan
luo ja avasi tuuletusruudun.
10.
— Tässä ei ole sitä. Älä ole huolissasi, minä tiedän, missä se on;
tässä se on, — sanoi Aljoša löydettyään huoneen toisesta nurkasta
Ivanin pesupöydän luota puhtaan, vielä kokoon taitetun ja
käyttämättömän pyyheliinan. Ivan katsoi omituisesti pyyheliinaan; oli
kuin hänen muistinsa olisi silmänräpäyksessä palannut.
— Kohta kaksitoista.
— Ei, ei, ei! — huudahti yhtäkkiä Ivan. — Se ei ollut uni! Hän oli,
hän istui täällä, tuolla sohvalla. Kun sinä koputit ikkunaan, paiskasin
minä häntä vastaan juomalasin… tämän… Odota, minä olen
ennenkin nukkunut, mutta tämä uni ei ole uni. Ja on ollut ennenkin.
Minä näen, Aljoša, nykyään unia… mutta ne eivät ole unia, vaan
tulevat valveilla ollessani, minä kävelen, puhun ja näen… mutta
nukun. Mutta hän istui tässä, hän oli, kas tuolla sohvalla… Hän on
hirveän tyhmä, Aljoša, hirveän tyhmä, — alkoi Ivan yhtäkkiä nauraa
ja rupesi astelemaan huoneessa.
— Niin, mutta hän on ilkeä. Hän pilkkasi minua. Hän oli röyhkeä,
Aljoša, — lausui Ivan vavahdellen loukkaantumisesta. — Mutta hän
panetteli minua, hän panetteli monessa suhteessa. Valehteli minusta
minulle itselleni vasten silmiä. »Oi, sinä menet suorittamaan hyveen
sankaritekoa, ilmoitat, että olet tappanut isäsi, että palvelija on sinun
yllyttämänäsi tappanut isäsi…»
— Hän sen sanoi, hän, ja hän tietää sen. »Sinä olet menossa
suorittamaan hyveen sankaritekoa, mutta hyveeseen sinä et
kuitenkaan usko — se juuri sinua suututtaa ja kiusaa, sentähden
sinä olet niin kostonhimoinen.» — Tätä hän sanoi minulle minusta
itsestäni, ja hän tietää, mitä puhuu…