Cardiovascular biomedical image analysis [Elektronische Ressource] : methods and applications / vorgelegt von Alexandru Paul Condurache

Sujets

• Informatik
• Gesundheit

Informations

• Publié par: universitat_zu_lubeck
• Publié le: 01 janvier 2006
• Nombre de lectures: 43
• Langue: English
• Poids de l'ouvrage: 20 Mo

Extrait

Institut für Signalverarbeitung und Prozessrechentechnik
Universität zu Lübeck
Direktor: Prof. Dr. Ing. Alfred Mertins
Cardiovascular Biomedical Image Analysis:
Methods and Applications
Inauguraldissertation
zur Erlangung der Doktorwürde (Dr. Ing.)
der Universität zu Lübeck
– Technisch Naturwissenschaftliche Fakultät –
vorgelegt von
Diplom Ingenieur
Alexandru Paul Condurache
aus Lübeck
Lübeck, im November 2006Referent: Priv. Doz. Dr. Ing. Erhardt Barth
Koreferent: Prof. Dr. Ing. Til Aach
Vorsitz des Prüfungsausschusses: Prof. Dr. Ing. Erik Maehle
Tag der mündlichen Prüfung: 23. Februar 2007
gezeichnet: Prof. Dr.rer.nat. Enno Hartmann
– Dekan der Technisch Naturwissenschaftlichen Fakultät der Universität zu Lübeck –Cardiovascular Biomedical Image Analysis: Methods
and Applications
Alexandru Paul CondurachePreface
My interest in seeing and thinking machines was ignited mainly during my high school years
by novels such as those written by Isaac Asimov and Arthur C. Clark, which however, found
the ground already prepared by Jules Verne. My study at the “Electronica” Faculty of the
“Polithenica” University of Bucharest was then the consequence. The passage from science
fiction to something more “earthly” happened during the last years as a student, while I was
1specializing in “Imagini, Forme si Inteligenta Artificiala (IFIA)” under the guidance of my
professor Vasile Buzuloiu and the group he lead at the “Laboratorul de Analiza si Prelucrarea
2Imaginilor (LAPI)” . From this group – to whom my gratitude is directed as a whole – I would
like to mention Dr. Mihai Ciuc and Dr. Constantin Vertan, which found the best way to make us
students the acquaintance with Image Processing and other related fields in a clear and enjoyable
way.
During my activity at LAPI I met Dr. Erhardt Barth, who gave me the possibility to pre
pare my diploma thesis at the Institute for Signal Processing of the then Medical University of
Lübeck, within a project related to vision based quality control. He has my gratitude and full
appreciation for being not only a teacher but also a very good friend.
During my time in Lübeck – an excellence center in medical technologies – I was first ac
quainted with medical image processing, an acquaintance which shaped my career from then on
trough the additional specialization year in medical imagery, again at the “Politehnica” Univer-
sity and then during my later stay as a research associate at the Institute for Signal Processing
(ISIP) of the University of Lübeck. This doctoral thesis has grown in a natural way from my in
terest in the field. It benefited largely from the physical proximity between ISIP and the Lübeck
campus of the University Clinic of Schleswig Holstein (UKSH), as well as from the innovative
and open spirit of its medical staff, from which I would like to mention those with whom I
collaborated more closely during the last years: Dr. Stephan Grzybowsky, Dr. Peter Radke, Dr.
Axel Kaiser, Dr. Martin Misfeld and especially Dr. Hans Günther Machens.
My special gratitude goes to Professor Til Aach, head of ISIP until 2004 and the main
“guiding light” during my time spent researching for this thesis. The numerous discussions
we had as well as his critical, well balanced points have been a constant source of scientific
inspiration just as much as his basketball skills, proven during the many games we’ve played
together, have helped establishing a relaxed, friendly and therefore – in my opinion – optimal
working atmosphere.
Towards my colleagues and friends at ISIP: Andre Folkers, Uli Hofmann, Christian Kier,
Kerstin Menne, Volker Metzler, Ingo Stuke and Daniel Toth goes my thanks not only for the
scientific debates we held together, but also for teaching me german and then enduring my
1Images Patterns and Artificial Intelligence
2The Image Processing and Analysis Laboratory (IPAL)endless speeches with stoicism. I also thank them for making me feel very well during my
entire stay in Germany until now.
For their support and understanding I thank my family: my parents Anca and Serban Con
durache who were there for me whenever I needed them and most specially to my wife Andreea
whose beautiful smile was always near to provide me with the emotional support I looked for.
I would like to end this Preface with my personal conclusion for these last five years: “Small
opportunities are often the beginning of great enterprises” – Demosthenes
To my FamilyContents
Abstract v
Zusammenfassung vii
1 Introduction 1
1.1 Biomedical cardiovascular image analysis . . . . . . . . . . . . . . . . . . . . 1
1.2 An enhanced catheter intervention . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Vessel segmentation and quantification in 2D projection images . . . . . . . . 3
1.4 Quality control for heart valves . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Machine vision for cardiovascular medicine: a tutorial 5
2.1 From human to machine vision . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 From photons to discrete signal . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.1 Image formation and detection with visible light . . . . . . . . . . . . 8
2.2.2 Digital imaging with X ray . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2.3 Image digitization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.2.4 noise and image distortions . . . . . . . . . . . . . . . . . . . . 32
2.3 Image processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.1 Image enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.2 Image analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.3.3 Morphological image processing . . . . . . . . . . . . . . . . . . . . . 39
2.4 Pattern recognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4.1 Statistical Classification . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.5 Issues of algorithm performance . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.5.1 Performance metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.5.2 Measuring in practice . . . . . . . . . . . . . . . . . . . 45
2.5.3 Designing a test to sustain a performance claim . . . . . . . . . . . . . 46
3 Enhanced catheter intervention 47
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.1.1 The coronary artery disease . . . . . . . . . . . . . . . . . . . . . . . 47
3.1.2 Percutaneous Transluminal Coronary Angioplasty . . . . . . . . . . . 48
3.1.3 Improved navigation in PTCA . . . . . . . . . . . . . . . . . . . . . . 53
3.1.4 Quantification of the Myocardial Blush . . . . . . . . . . . . . . . . . 55
3.2 A dynamic roadmap for enhanced navigation in PTCA . . . . . . . . . . . . . 55
3.2.1 Analysis of contrasted images . . . . . . . . . . . . . . . . . . . . . . 56
iii CONTENTS
3.2.2 Sequence matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.2.3 Segmentation and registration of surgical tools . . . . . . . . . . . . . 88
3.3 Quantitative analysis of the myocardial blush . . . . . . . . . . . . . . . . . . 95
3.3.1 A heart motion compensated ROI . . . . . . . . . . . . . . . . . . . . 96
3.3.2 Robust analysis of the myocardial blush . . . . . . . . . . . . . . . . . 103
3.4 Conclusions and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.4.1 The dynamic roadmap . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.4.2 Automatic estimation of the MBG . . . . . . . . . . . . . . . . . . . . 110
4 Vessel segmentation in 2D projection images 113
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.1.1 Vascular imaging methods . . . . . . . . . . . . . . . . . . . . . . . . 113
4.1.2 Vessel segmentation as a pattern recognition problem . . . . . . . . . . 114
4.1.3 A framework for vessel segmentation . . . . . . . . . . . . . . . . . . 116
4.2 Vessel enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.2.1 Background attenuation . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.2.2 Vessel augmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.2.3 Pixel based multidimensional description of vessel and background . . 126
4.2.4 Vessel enhancement: Experiments and discussion . . . . . . . . . . . . 126
4.2.5 Vessel Conclusions . . . . . . . . . . . . . . . . . . . . 129
4.3 Automatic vessel segmentation in 2D projection images . . . . . . . . . . . . . 131
4.3.1 Segmentation by thresholding . . . . . . . . . . . . . . . . . . . . . . 131
4.3.2 Hysteresis segmentation . . . . . . . . . . . . . . . . . . . . . . . . . 136
4.3.3 Automatic vessel segmentation: Experiments and results . . . . . . . . 151
4.3.4 vessel se Discussion . . . . . . . . . . . . . . . 157
4.3.5 Automatic vessel segmentation: Conclusions . . . . . . . . . . . . . . 161
4.4 A semi automatic system for the analysis of angiogenesis in skin transplant
microangiograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.4.1 A tool for the analysis of angiogenesis: Introduction . . . . . . . . . . 163
4.4.2 A tool for the analysis of The segmentation of vessels . . 165
4.4.3 A tool for the of Description of vessel