Cellular image classification

142 Pages·2017·2.686 MB·English

by  Lin, Feng, Wu, Xingkun, Xu, Xiang

Xiang Xu · Xingkun Wu
Feng Lin
Cellular
Image
Classification
fi
Cellular Image Classi cation
Xiang Xu Xingkun Wu
(cid:129)
Feng Lin
fi
Cellular Image Classi cation
123
XiangXu FengLin
Schoolof Computer Engineering Schoolof Computer Engineering
NanyangTechnological University NanyangTechnological University
Singapore Singapore
Xingkun Wu
Zhejiiang University
Hangzhou, Zhejiang
China
ISBN978-3-319-47628-5 ISBN978-3-319-47629-2 (eBook)
DOI 10.1007/978-3-319-47629-2
LibraryofCongressControlNumber:2016955784
©SpringerInternationalPublishingAG2017
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor
foranyerrorsoromissionsthatmayhavebeenmade.
Printedonacid-freepaper
ThisSpringerimprintispublishedbySpringerNature
TheregisteredcompanyisSpringerInternationalPublishingAG
Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland
Preface
This book introduces new techniques for cellular image feature extraction, pattern
recognitionandclassification.Weusetheantinuclearantibodies(ANAs)inpatient
serum as the subjects and the indirect immunofluorescence (IIF) technique as the
imagingprotocoltoillustratetheapplicationsofthedescribedmethods.Throughout
the book, we provide evaluations for our proposed methods on two publicly
available human epithelial (HEp-2) cell datasets: ICPR2012 dataset from the
ICPR’12 HEp-2 cell classification contest and ICIP2013 training dataset from the
ICIP’13 Competition on cells classification by fluorescent image analysis.
First, the reading of imaging results is significantly influenced by one’s quali-
ficationandreadingsystems,causinghighintra- andinter-laboratoryvariance.We
describe a low-order LP21 fiber mode for optical single cell manipulation and
imagingstainingpatternsofHEp-2cells.Afocusedfour-lobedmodedistributionis
stableandeffectiveinopticaltweezerapplications,includingselectivecellpick-up,
pairing,groupingorseparation,aswellasrotationofcelldimersandclusters.Both
translational dragging force and rotational torque in the experiments are in good
accordance with our theoretical model. With a simple all-fiber configuration, and
low peak irradiation to targeted cells, instrumentation of our optical chuck tech-
nology will provide a powerful tool in the ANA-IIF laboratories. We focus on the
optical, mechanical and computing systems for the clinical trials. Computer pro-
grams for GUI and control of the optical tweezers are also discussed.
Next, we introduce the Bag-of-Words (BoW) framework which is one of the
most successful image representations. To reduce the inevitable information loss
causedbycodingprocess,westudyalinearlocaldistancecoding(LLDC)method.
The LLDC method transforms original local feature to more discriminative local
distance vector by searching for local neighbors of the local feature in the
class-specificmanifolds.Thenweencodeandpoolthelocaldistancevectorstoget
salient image representation. Combined with the traditional coding methods, this
method achieves a higher classification accuracy.
Then, we study a rotation invariant textural feature of pairwise local ternary
patterns with spatial rotation invariant (PLTP-SRI). It is invariant to image rota-
tions, meanwhile it is robust to noise and weak illumination. By adding spatial
v
vi Preface
pyramid structure, this method captures spatial layout information. While the
proposed PLTP-SRI feature extracts local feature, the BoW framework builds a
global image representation. It is reasonable to combine them together to achieve
impressiveclassificationperformance,asthecombinedfeaturetakestheadvantages
of the two kinds offeatures in different aspects.
Finally, we design a co-occurrence differential texton (CoDT) feature to repre-
sent the local image patches of HEp-2 cells. The CoDT feature reduces the infor-
mationlossbyignoringthequantizationwhileitutilizesthespatialrelationsamong
thedifferential micro-textonfeature. Thusitcanincreasethediscriminative power.
We build a generative model to adaptively characterize the CoDT feature space
of the training data. Furthermore, we exploit a discriminant representation for the
HEp-2 cell images based on the adaptive partitioned feature space. Therefore, the
resulting representation is adapted to the classification task. By cooperating with
linear support vector machine (SVM) classifier, our framework can exploit the
advantages of both generative and discriminative approaches for cellular image
classification.
Themonographiswrittenforthoseresearcherswhowouldliketodeveloptheir
ownprograms,andtheworkingMATLABcodesareincludedforalltheimportant
algorithms presented. It can also be used asa reference book for graduate students
and senior undergraduates in the area of biomedical imaging, image feature
extraction,patternrecognitionandclassification.Academics,researchers,andmany
others will find this to be an exceptional resource.
Enjoy the read.
Singapore Xiang Xu
China Xingkun Wu
Singapore Feng Lin
August 2016
Contents
1 Introduction.... .... .... ..... .... .... .... .... .... ..... .... 1
1.1 Background .... .... ..... .... .... .... .... .... ..... .... 1
1.1.1 Clinical Problems: A Case Study on Autoimmune
Diseases . .... ..... .... .... .... .... .... ..... .... 1
1.1.2 Cellular Imaging: A Case Study on Indirect
Immunofluorescence . .... .... .... .... .... ..... .... 3
1.2 Computer-Aided Diagnosis.. .... .... .... .... .... ..... .... 6
1.3 Experimental Datasets in the Book.... .... .... .... ..... .... 8
1.3.1 The ICPR2012 Dataset ... .... .... .... .... ..... .... 8
1.3.2 The ICIP2013 Training Dataset. .... .... .... ..... .... 10
1.4 Structure of the Chapters ... .... .... .... .... .... ..... .... 10
References.. .... .... .... ..... .... .... .... .... .... ..... .... 12
2 Fundamentals .. .... .... ..... .... .... .... .... .... ..... .... 15
2.1 Optical Systems for Cellular Imaging.. .... .... .... ..... .... 15
2.1.1 Laser Scanning Confocal Microscope.... .... ..... .... 16
2.1.2 Multi-photon Fluorescence Imaging . .... .... ..... .... 20
2.1.3 Total Internal Reflection Fluorescence Microscope... .... 22
2.1.4 Near-Field Scanning Optical Microscopy Imaging
Technology... ..... .... .... .... .... .... ..... .... 25
2.1.5 Optical Coherence Tomography Technology... ..... .... 29
2.2 Feature Extraction.... ..... .... .... .... .... .... ..... .... 31
2.2.1 Low-Level Features.. .... .... .... .... .... ..... .... 31
2.2.2 Mid-Level Features.. .... .... .... .... .... ..... .... 38
2.3 Classification ... .... ..... .... .... .... .... .... ..... .... 39
2.3.1 Support Vector Machine.. .... .... .... .... ..... .... 39
2.3.2 Nearest Neighbor Classifier.... .... .... .... ..... .... 40
References.. .... .... .... ..... .... .... .... .... .... ..... .... 41
vii
viii Contents
3 Optical Systems for Cellular Imaging .... .... .... .... ..... .... 45
3.1 Introduction .... .... ..... .... .... .... .... .... ..... .... 46
3.2 Optical Tweezer. .... ..... .... .... .... .... .... ..... .... 47
3.2.1 Introduction to Optical Tweezers ... .... .... ..... .... 47
3.2.2 Gradient and Scattering Force of Optical Tweezers... .... 48
3.2.3 Three-Dimensional Optical Trap.... .... .... ..... .... 49
3.3 Low-Order Fiber Mode LP .... .... .... .... .... ..... .... 51
21
3.3.1 Fiber Mode Coupling Theory .. .... .... .... ..... .... 51
3.3.2 Analysis of Field Distribution in Optical Fiber. ..... .... 53
3.3.3 Solution to LP Mode ... .... .... .... .... ..... .... 55
21
3.3.4 Selective Excitation of LP Mode .. .... .... ..... .... 56
21
3.3.5 The Twisting and Bending Characteristics
of LP Mode. ..... .... .... .... .... .... ..... .... 58
21
3.3.6 Why LP Mode? ... .... .... .... .... .... ..... .... 60
21
3.4 Optical Tweezer Using Focused LP Mode. .... .... ..... .... 61
21
3.4.1 Fiber Axicons. ..... .... .... .... .... .... ..... .... 61
3.4.2 Cell Manipulation... .... .... .... .... .... ..... .... 66
3.5 Modeling of Optical Trapping Force .. .... .... .... ..... .... 68
3.5.1 Force Analysis of Mie Particles in Optical Trap..... .... 69
3.5.2 Gaussian Beam..... .... .... .... .... .... ..... .... 72
3.5.3 Simulation of Light Force on Mie Particle .... ..... .... 73
3.6 Summary .. .... .... ..... .... .... .... .... .... ..... .... 77
References.. .... .... .... ..... .... .... .... .... .... ..... .... 78
4 Image Representation with Bag-of-Words. .... .... .... ..... .... 81
4.1 Introduction .... .... ..... .... .... .... .... .... ..... .... 81
4.2 Coding .... .... .... ..... .... .... .... .... .... ..... .... 83
4.2.1 Vector Quantization . .... .... .... .... .... ..... .... 84
4.2.2 Soft Assignment Coding.. .... .... .... .... ..... .... 84
4.2.3 Locality-Constrained Linear Coding . .... .... ..... .... 85
4.3 Pooling.... .... .... ..... .... .... .... .... .... ..... .... 86
4.4 Summary .. .... .... ..... .... .... .... .... .... ..... .... 86
References.. .... .... .... ..... .... .... .... .... .... ..... .... 86
5 Image Coding .. .... .... ..... .... .... .... .... .... ..... .... 89
5.1 Introduction .... .... ..... .... .... .... .... .... ..... .... 89
5.2 Linear Local Distance Coding Method. .... .... .... ..... .... 90
5.2.1 Distance Vector..... .... .... .... .... .... ..... .... 91
5.2.2 Local Distance Vector.... .... .... .... .... ..... .... 92
5.2.3 The Algorithm Framework .... .... .... .... ..... .... 93
5.3 Experiments and Analyses .. .... .... .... .... .... ..... .... 94
5.3.1 Experiment Setup ... .... .... .... .... .... ..... .... 95
5.3.2 Experimental Results on the ICPR2012 Dataset ..... .... 96
Contents ix
5.3.3 Experimental Results on the ICIP2013
Training Dataset .... .... .... .... .... .... ..... .... 98
5.3.4 Discussion.... ..... .... .... .... .... .... ..... .... 99
5.4 Summary .. .... .... ..... .... .... .... .... .... ..... .... 102
References.. .... .... .... ..... .... .... .... .... .... ..... .... 102
6 Encoding Image Features. ..... .... .... .... .... .... ..... .... 105
6.1 Introduction .... .... ..... .... .... .... .... .... ..... .... 105
6.2 Encoding Rotation Invariant Features of Images.. .... ..... .... 107
6.2.1 Pairwise LTPs with Spatial Rotation Invariant . ..... .... 107
6.2.2 Encoding the SIFT Features with BoW Framework .. .... 110
6.3 Experiments and Analyses .. .... .... .... .... .... ..... .... 111
6.3.1 Experiment Setup ... .... .... .... .... .... ..... .... 111
6.3.2 Experimental Results on the ICPR2012 Dataset ..... .... 112
6.3.3 Experimental Results on the ICIP2013 Training
Dataset .. .... ..... .... .... .... .... .... ..... .... 113
6.3.4 Discussion.... ..... .... .... .... .... .... ..... .... 115
6.4 Summary .. .... .... ..... .... .... .... .... .... ..... .... 117
References.. .... .... .... ..... .... .... .... .... .... ..... .... 117
7 Defining Feature Space for Image Classification.... .... ..... .... 119
7.1 Introduction .... .... ..... .... .... .... .... .... ..... .... 119
7.2 Adaptive Co-occurrence Differential Texton Space
for Classification. .... ..... .... .... .... .... .... ..... .... 120
7.2.1 Co-occurrence Differential Texton... .... .... ..... .... 120
7.2.2 Adaptive CoDT Feature Space . .... .... .... ..... .... 123
7.2.3 HEp-2 Cell Image Representation
in the Adaptive CoDT Feature Space .... .... ..... .... 124
7.3 Experiments and Analyses .. .... .... .... .... .... ..... .... 127
7.3.1 Experiment Setup ... .... .... .... .... .... ..... .... 127
7.3.2 Experimental Results on the ICPR2012 Dataset ..... .... 128
7.3.3 Experimental Results on the ICIP2013
Training Dataset .... .... .... .... .... .... ..... .... 129
7.3.4 Discussion.... ..... .... .... .... .... .... ..... .... 130
7.4 Summary .. .... .... ..... .... .... .... .... .... ..... .... 132
References.. .... .... .... ..... .... .... .... .... .... ..... .... 132
8 Conclusions and Perspectives... .... .... .... .... .... ..... .... 135
8.1 Major Techniques Developed in the Book.. .... .... ..... .... 135
8.2 Directions and Future Work. .... .... .... .... .... ..... .... 136
References.. .... .... .... ..... .... .... .... .... .... ..... .... 137
Chapter 1
Introduction
Abstract In this chapter, we introduce the background of our study, followed by
the research motivations and objectives. Then, we introduce the publicly available
datasetsusedtoevaluateourproposedmethods.Finally,wesummarizeourcontri-
butionsandgivethestructureofthereportintheend.
1.1 Background
1.1.1 ClinicalProblems:ACaseStudyonAutoimmune
Diseases
Autoimmunediseases(AD)occurwhentheimmunesystemmistakenlyattacksand
destroyshealthybodytissues,whichcanbeorgan-specificorsystemic.Depending
onthetype,anautoimmunediseasecanaffectoneormanydifferenttypesofbody
tissue. It can also cause abnormal organ growth and changes in organ function.
They are prevalent diseases affecting a large number of human beings. There are
morethaneightydifferenttypesofADsuchasrheumatoidarthritis,systemiclupus
erythematosus, scleroderma and autoimmune hepatitis. The common autoimmune
diseasesareintroducedasfollows:
(cid:129) rheumatoidarthritis:chronicinflammationofjointsandsurroundingtissues;
(cid:129) systemiclupuserythematosus:affectsskin,joints,kidneys,brainandotherorgans;
(cid:129) scleroderma:aconnectivetissuediseasethatcauseschangesinskin,bloodvessels,
muscles,andinternalorgans;
(cid:129) sjogren’ssyndrome:destroystheglandsthatproducetearsandsalivacausingdry
eyesandmouth;mayaffectkidneysandlungs;
(cid:129) autoimmunehepatitis:causesinflammationandliverdamage;
(cid:129) perniciousanemia:decreaseinredbloodcellscausedbyinabilitytoabsorbvitamin
B-12;
(cid:129) vitiligo:whitepatchesontheskincausedbylossofpigment;
(cid:129) psoriasis:askinconditionthatcausesrednessandirritationaswellasthick,flaky,
silver-whitepatches;
©SpringerInternationalPublishingAG2017 1
X.Xuetal.,CellularImageClassification,
DOI10.1007/978-3-319-47629-2_1