Dermoscopy Image Analysis

458 Pages·2015·58.739 MB·\458

by  M. Emre Celebi (Editor), Teresa Mendonca (Editor), Jorge S. Marques (Editor)

Biomedical Science
Celebi
Mendonça
Marques
“… demonstrates the significant advancements in dermoscopy image analysis
techniques, involving image-preprocessing, lesion segmentation (or border detection),
feature extraction, pattern analysis, lesion classification, and database construction. The
approaches described are state of the art and the details are sine qua non. … a good
addition on the dermatologist and researcher’s bookshelf, and pilots further research in
this field.”
—Fengying Xie, Image Processing Center, Beihang University, Beijing, China
“… a comprehensive description of computerized image analysis that is definitely useful
for researchers in this field as a starting point for future developments.”
—Giuseppe Argenziano, Dermatology Unit, Second University of Naples, Italy
“… collect[s] high-quality research articles on dermoscopy image analysis. … presents
the latest state-of-the-art techniques for classifying benign and malignant skin lesions.
A very interesting point that emerged from the literature review, as one author points
out, is that all these computer-aided diagnosis (CAD) systems offer invaluable help
to the general practitioner (GP) in an initial diagnosis of the lesions. The high level of
performance of these systems can thus provide an accurate evaluation of skin lesions
without having to refer to a dermatologist at this initial stage.”
—Lucia Ballerini, University of Edinburgh, Scotland
“… deals with a very important and hot topic that attracts the interest of many
researchers and doctors in the scientific community nowadays. … well written and clear.
It provides a good overview of the dermoscopy image analysis field, and it is useful for
engineers and computer scientists interested in developing similar applications, since it
includes in-depth technical descriptions.”
—Ilias Maglogiannis, University of Piraeus, Greece
“… I definitely would like to have this book [on] my shelf. … a valuable resource for
researchers and graduate students of computer vision and medical image analysis
interested in skin cancer detection methods for dermoscopic images. It covers
fundamentals of the area by providing thorough treatments of the theory and the
concepts while making the material accessible to the reader with examples that nicely
illustrate the concepts.”
—Jacob Scharcanski, Universidade Federal do Rio Grande do Sul, Brazil

“… provides a comprehensive, well-rounded coverage on the factors, challenges, and
state-of-the-art solutions to the very difficult problem of dermascopy image analysis.
… a good reference to have in the libraries of researchers in this field.”
—Alexander Wong, University of Waterloo, Ontario, Canada
K23910
ISBN: 978-1-4822-5326-9
90000
9 781482 253269
K23910_COVER.indd 1 8/14/15 4:00 PM
DERMOSCOPY
IMAGE ANALYSIS
T&F Cat #K23910 — K23910 C000 — page i — 8/14/2015 — 10:48
Digital Imaging and Computer Vision Series
Series Editor
Rastislav Lukac
Foveon, Inc./Sigma Corporation
San Jose, California, U.S.A.
Dermoscopy Image Analysis, by M. Emre Celebi, Teresa Mendonça, and Jorge S. Marques
Semantic Multimedia Analysis and Processing, by Evaggelos Spyrou, Dimitris Iakovidis,
and Phivos Mylonas
Microarray Image and Data Analysis: Theory and Practice, by Luis Rueda
Perceptual Digital Imaging: Methods and Applications, by Rastislav Lukac
Image Restoration: Fundamentals and Advances, by Bahadir Kursat Gunturk and Xin Li
Image Processing and Analysis with Graphs: Theory and Practice, by Olivier Lézoray
and Leo Grady
Visual Cryptography and Secret Image Sharing, by Stelvio Cimato and Ching-Nung Yang
Digital Imaging for Cultural Heritage Preservation: Analysis, Restoration, and
Reconstruction of Ancient Artworks, byFilippo Stanco, Sebastiano Battiato, and Giovanni Gallo
Computational Photography: Methods and Applications, by Rastislav Lukac
Super-Resolution Imaging, by Peyman Milanfar
T&F Cat #K23910 — K23910 C000 — page ii — 8/14/2015 — 10:48
DERMOSCOPY
IMAGE ANALYSIS
Edited by
M. Emre Celebi
Louisiana State University, Shreveport, USA
Teresa Mendonça
University of Porto, Portugal
Jorge S. Marques
Instituto Superior Tecnico, Lisboa, Portugal
Boca Raton London New York
CRC Press is an imprint of the
Taylor & Francis Group, an informa business
T&F Cat #K23910 — K23910 C000 — page iii — 8/14/2015 — 10:48
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not
warrant the accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB® soft-
ware or related products does not constitute endorsement or sponsorship by The MathWorks of a particular
pedagogical approach or particular use of the MATLAB® software.
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2016 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
Version Date: 20150608
International Standard Book Number-13: 978-1-4822-5327-6 (eBook - PDF)
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, transmit-
ted, 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 (http://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.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Contents
Preface.......................................................................................................vii
Editors........................................................................................................xi
Contributors............................................................................................xiii
Chapter 1 Toward a Robust Analysis of Dermoscopy Images
Acquired under Different Conditions.....................................1
Catarina Barata, M. Emre Celebi, and Jorge S. Marques
Chapter 2 A Bioinspired Color Representation for Dermoscopy
Image Analysis.....................................................................23
Ali Madooei and Mark S. Drew
Chapter 3 Where’s the Lesion?: Variability in Human and
Automated Segmentation of Dermoscopy Images of
Melanocytic Skin Lesions.....................................................67
Federica Bogo, Francesco Peruch, Anna Belloni Fortina,
and Enoch Peserico
Chapter 4 A State-of-the-Art Survey on Lesion Border
Detection in Dermoscopy Images.........................................97
M.EmreCelebi,QuanWen,HitoshiIyatomi,KouheiShimizu,
Huiyu Zhou, and Gerald Schaefer
Chapter 5 Comparison of Image Processing Techniques for
Reticular Pattern Recognition in Melanoma Detection.....131
Jose Luis Garc´ıa Arroyo and Begon˜a Garc´ıa Zapirain
Chapter 6 Global Pattern Classification in Dermoscopic Images........183
Aurora Sa´ez, Carmen Serrano, and Begon˜a Acha
Chapter 7 Streak Detection in Dermoscopic Color Images
Using Localized Radial Flux of Principal Intensity
Curvature...........................................................................211
Hengameh Mirzaalian, Tim K. Lee, and Ghassan Hamarneh
v
T&F Cat #K23910 — K23910 C000 — page v — 8/14/2015 — 10:48
vi Contents
Chapter 8 Dermoscopy Image Assessment Based on Perceptible
Color Regions.....................................................................231
Gunwoo Lee, Onseok Lee, Jaeyoung Kim, Jongsub Moon,
and Chilhwan Oh
Chapter 9 Improved Skin Lesion Diagnostics for General
Practice by Computer-Aided Diagnostics..........................247
Kajsa Møllersen, Maciel Zortea, Kristian Hindberg,
Thomas R. Schopf, Stein Olav Skrøvseth, and
Fred Godtliebsen
Chapter 10 Accurate and Scalable System for Automatic
Detection of Malignant Melanoma.....................................293
ManiAbedini,QiangChen,NoelC.F.Codella,RahilGarnavi,
and Xingzhi Sun
Chapter 11 Early Detection of Melanoma in Dermoscopy of
Skin Lesion Images by Computer Vision–Based System....345
Hoda Zare and Mohammad Taghi Bahreyni Toossi
Chapter 12 From Dermoscopy to Mobile Teledermatology...................385
Lu´ıs Rosado, Maria Joa˜o M. Vasconcelos, Rui Castro,
and Jo˜ao Manuel R. S. Tavares
Chapter 13 PH2: A Public Database for the Analysis of
Dermoscopic Images...........................................................419
Teresa F. Mendonc¸a, Pedro M. Ferreira, Andr´e R. S. Marc¸al,
Catarina Barata, Jorge S. Marques, Joana Rocha,
and Jorge Rozeira
Index........................................................................................................441
T&F Cat #K23910 — K23910 C000 — page vi — 8/14/2015 — 10:48
Preface
Malignant melanoma is one of the most rapidly increasing cancers in the
world. Invasive melanoma alone has an estimated incidence of 73,870 and an
estimatedtotalof9940deathsintheUnitedStatesin2015[1].Earlydiagnosis
is particularly important since melanoma can be cured with a simple excision
if detected early.
In the past, the primary form of diagnosis for melanoma has been unaided
clinicalexamination.Inrecentyears,dermoscopyhasprovedvaluableinvisu-
alizingthemorphologicalstructuresinpigmentedlesions.However,ithasalso
beenshownthatdermoscopyisdifficulttolearnandsubjective.Therefore,the
development of automated image analysis techniques for dermoscopy images
has gained importance.
The goal of this book is to summarize the state of the art in the compu-
terized analysis of dermoscopy images and provide future directions for this
exciting subfield of medical image analysis. The intended audience includes
researchers and practicing clinicians, who are increasingly using digital
analytic tools.
The book opens with two chapters on preprocessing. In “Toward a Robust
AnalysisofDermoscopyImagesAcquiredunderDifferentConditions,”Barata
etal.investigatetheinfluenceofcolornormalizationonclassificationaccuracy.
Theauthorsinvestigatethreecolorconstancyalgorithms,namely,grayworld,
max-RGB,andshadesofgray,anddemonstratesignificantgainsinsensitivity
and specificity on a heterogeneous set of images. In “A Bioinspired Color
RepresentationforDermoscopyImageAnalysis,”MadooeiandDrewpropose
a new color space that highlights the distribution of underlying melanin and
hemoglobincolorpigments.Theadvantageofthisnewcolorrepresentation,in
additiontoitsbiologicalunderpinnings,liesinitsattenuationoftheeffectsof
confounding factors such as light color, intensity falloff, shading, and camera
characteristics. The authors demonstrate that the new color space leads to
more accurate classification and border detection results.
Thebookcontinueswithtwochaptersonborderdetection(segmentation).
In “Where’s the Lesion? Variability in Human and Automated Segmentation
of Dermoscopy Images of Melanocytic Skin Lesions,” Bogo et al. examine
the extent of agreement among dermatologist-drawn borders and that among
dermatologist-drawnbordersandautomaticallydeterminedones.Theauthors
conclude that state-of-the-art border detection algorithms can achieve a level
of agreement that is only slightly lower than the level of agreement among
experienced dermatologists themselves. In “A State-of-the-Art Survey on
LesionBorderDetectioninDermoscopyImages,”Celebietal.presentacom-
prehensive overview of 50 published border detection methods. The authors
vii
T&F Cat #K23910 — K23910 C000 — page vii — 8/14/2015 — 10:48
viii Preface
reviewpreprocessing,segmentation,andpostprocessingaspectsofthesemeth-
ods and discuss performance evaluation issues. They also propose guidelines
for future studies in automated border detection.
The book continues with four chapters on feature extraction. In
“Comparison of Image Processing Techniques for Reticular Pattern Recog-
nition in Melanoma Detection,” Garc´ıa Arroyo and Garc´ıa Zapirain present
an in-depth overview of the state of the art in the extraction of pigment net-
works from dermoscopy images. The authors give a detailed explanation of
20 selected methods and then compare them with respect to various crite-
ria, including the number and diagnostic distribution of the images used for
validation and the numerical results obtained in terms of sensitivity, speci-
ficity,andaccuracy.In“GlobalPatternClassificationinDermoscopicImages,”
Sa´ez et al. present an overview of six methods for extracting the global
patterns (namely, reticular, globular, cobblestone, homogeneous, starburst,
parallel, multicomponent, and lacunar patterns) as defined in the pattern
analysis diagnostic scheme. The authors first illustrate each pattern and then
describe the automated methods designed for extracting these patterns. The
chapter concludes with a critical discussion of global pattern extraction. In
“Streak Detection in Dermoscopic Color Images Using Localized Radial Flux
of Principal Intensity Curvature,” Mirzaalian et al. present an automated
method for detecting streaks based on the concept of quaternion tubularness
andnonlinearsupportvectormachineclassification.Theauthorsdemonstrate
the performance of their feature extraction method on 99 images from the
EDRA atlas. Finally, in “Dermoscopy Image Assessment Based on Percep-
tible Color Regions,” Lee et al. present a method for detecting perceptually
significantcolorsindermoscopyimages.Theauthorsfirstpartitiontheimage
into 27 color regions by dividing each of the red, green, and blue channels
intothreelevelsofbrightnessusingamultithresholdingalgorithm.Theythen
extract various color features from these regions. The classification perfor-
mance of these features is demonstrated on 150 images obtained from the
Korea University Guro Hospital.
The book continues with four chapters on classification. In “Improved
Skin Lesion Diagnostics for General Practice by Computer-Aided Diagnos-
tics,” Møllersen et al. present a computer-aided diagnosis (CAD) system for
melanomasthatfeaturesaninexpensiveacquisitiontool,clinicallymeaningful
features, and interpretable classification feedback. The authors evaluate their
systemon206imagesacquiredattwosites.In“AccurateandScalableSystem
for Automatic Detection of Malignant Melanoma,” Abedini et al. present a
comprehensiveliteraturereviewonCADsystemsformelanomas.Theauthors
then propose a highly scalable CAD system implemented in the MapReduce
framework and demonstrate its performance on approximately 3000 images
obtained from two sources. In “Early Detection of Melanoma in Dermoscopy
of Skin Lesion Images by a Computer Vision–Based System,” Zare and
Toossi present a novel CAD system for melanomas that involves hair
detection/removal based on edge detection, thresholding, and inpainting;
T&F Cat #K23910 — K23910 C000 — page viii — 8/14/2015 — 10:48
Preface ix
border detection using region-based active contours; extraction of various
low-level features; feature selection using the t-test; and classification using a
neural network classifier. The authors demonstrate the performance of their
systemon322imagesobtainedfromtwodermoscopyatlases.Finally,in“From
Dermoscopy to Mobile Teledermatology,” Rosado et al. discuss telemedicine
aspects of dermatology. The authors first present an overview of dermato-
logical image databases. They then discuss the challenges involved in the
preprocessing of clinical skin lesion images acquired with mobile devices and
describe a patient-oriented system for analyzing such images. Finally, they
conclude with a comparative review of smart phone-adaptable dermoscopes.
Achaptertitled“PH2:APublicDatabasefortheAnalysisofDermoscopic
Images” by Mendonc¸a et al. completes the book. The authors present a pub-
liclyavailabledatabaseofdermoscopyimages,whichcontains200high-quality
images along with their medical annotations. This database can be used as
ground truth in various dermoscopy image analysis tasks, including prepro-
cessing, border detection, feature extraction, and classification. The authors
also describe some of their projects that made use of this database.
Aseditors,wehopethatthisbookoncomputerizedanalysisofdermoscopy
imageswilldemonstratethesignificantprogressthathasoccurredinthisfield
inrecentyears.Wealsohopethatthedevelopmentsreportedinthisbookwill
motivate further research in this exciting field.
REFERENCE
1. R.L.Siegel,K.D.Miller,andA.Jemal,“CancerStatistics,2015,”CA:ACancer
Journal for Clinicians,vol.65,no.1,pp.5–29,2015.
M. Emre Celebi
Louisiana State University
Shreveport, Louisiana
Teresa F. Mendonc¸a
Universidade do Porto
Porto, Portugal
Jorge S. Marques
Instituto Superior T´ecnico
Lisbon, Portugal
T&F Cat #K23910 — K23910 C000 — page ix — 8/14/2015 — 10:48