ebook img

Breath Analysis for Medical Applications PDF

312 Pages·2017·10.155 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Breath Analysis for Medical Applications

David Zhang · Dongmin Guo Ke Yan Breath Analysis for Medical Applications Breath Analysis for Medical Applications ⋅ David Zhang Dongmin Guo Ke Yan Breath Analysis for Medical Applications 123 DavidZhang Ke Yan Biometrics Research Centre National Institute ofHealth TheHong Kong Polytechnic University Bethesda Hong Kong USA China DongminGuo WakeForest University Winston-Salem, NC USA ISBN978-981-10-4321-5 ISBN978-981-10-4322-2 (eBook) DOI 10.1007/978-981-10-4322-2 LibraryofCongressControlNumber:2017939534 ©SpringerNatureSingaporePteLtd.2017 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 for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations. Printedonacid-freepaper ThisSpringerimprintispublishedbySpringerNature TheregisteredcompanyisSpringerNatureSingaporePteLtd. Theregisteredcompanyaddressis:152BeachRoad,#21-01/04GatewayEast,Singapore189721,Singapore Preface Somecomponentsinhumanbreathhavebeenproventobeassociatedwithcertain diseases and the concentration of these components is linked to disease status. Recently, breath signal diagnosis has attracted increasing research interests. Many kinds of breath signal acquisition systems and breath signal processing methods havebeenreported.However,there are still alot ofchallenging workstobedone, for example, how to acquire breathsignalin a fast, accurate, and informative way, howtopreprocess thebreathsignaltorule outtheoutliersandincreasethequality of the signal, and how to extract efficient features and find proper classifiers for breath diagnosis. This book focuses on these challenging issues. Novel breath signal acquisition systems based on multiple breath sensors were described first. In order to collect samples effectively, we developed a sample acquisition system with sensor fusion technology. To detect the drift of breath signals, we provided optimized prepro- cessing frameworks, such as using transfer samples and regression models. To represent breath signals completely, we discovered different types of breath signal features, such as spatial feature, frequency feature, deep learning feature, etc. Moreover, we also provided many effective algorithms for breath signal classifi- cation and recognition, such as curve-fitting models and sparse representation classification. All of the technologies, algorithms, and medical application cases described in this book were applied in our research work and have proven to be effective in breathsignalanalysis.First,thisbookpresentsacomprehensiveintroductiononthe useful techniques of breath signal acquisition methods using different kinds of chemical sensors, cooperated with the optimized selection and fusion acquisition scheme.Then,thisbookintroducestheeffectivepreprocessingapproaches,suchas driftremovingandfeatureextractionmethods.Moreover,theclassificationmethods used as case studies are also provided. Finally, this book provides discussions and concluding remarks to indicate some promising directions on the studies and medical applications of computerized breath diagnosis. This book will benefit the researchers, professionals, graduate and postgraduate students working in the field v vi Preface of breath sample diagnosis, signal processing, pattern recognition,biometrics, etc. This book will also be very useful for interdisciplinary research. Our team has been working on the breath analysis research on computational TCMdiagnosisover10years.Underthegrantsupport(GrantNo.61332011)from NationalNaturalScienceFoundationofChina(NSFC)andHongKongPolytechnic University,wehadstartedourstudiesonthistopic.Theauthorswouldliketothank Dr. Zhaotian Zhang, Dr. Xiaoyun Xiong, and Dr. Ke Liu from NSFC for their consistent support to our research work. Hong Kong, China David Zhang Winston-Salem, USA Dongmin Guo Bethesda, USA Ke Yan February 2017 Contents Part I Background 1 Introduction .... .... .... ..... .... .... .... .... .... ..... .. 3 1.1 Background and Motivation. .... .... .... .... .... ..... .. 3 1.1.1 Why Is Breath Analysis Used in Disease Diagnosis? .. ..... .... .... .... .... .... ..... .. 4 1.1.2 Why Should Breath Analysis System Be Developed?.... .... .... .... .... .... ..... .. 4 1.1.3 Why Should Specific Algorithms Be Designed for Breath Analysis?.... .... .... .... .... ..... .. 5 1.2 Relative Technologies ..... .... .... .... .... .... ..... .. 7 1.3 Outline of the Work.. ..... .... .... .... .... .... ..... .. 7 References .. .... .... .... ..... .... .... .... .... .... ..... .. 9 2 Literature Review.... .... ..... .... .... .... .... .... ..... .. 11 2.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 11 2.2 Development of Breath Analysis . .... .... .... .... ..... .. 13 2.3 Breath Analysis by GC .... .... .... .... .... .... ..... .. 14 2.3.1 Lung Cancer. ..... .... .... .... .... .... ..... .. 14 2.3.2 Lipid Peroxidation . .... .... .... .... .... ..... .. 15 2.3.3 Renal Diseases .... .... .... .... .... .... ..... .. 16 2.3.4 Liver Diseases..... .... .... .... .... .... ..... .. 17 2.3.5 Breast Cancer..... .... .... .... .... .... ..... .. 17 2.3.6 Diabetes .... ..... .... .... .... .... .... ..... .. 17 2.3.7 Pulmonary Tuberculosis . .... .... .... .... ..... .. 18 2.3.8 Summary ... ..... .... .... .... .... .... ..... .. 18 2.4 Breath Analysis by E-Nose . .... .... .... .... .... ..... .. 19 2.5 Summary.. .... .... ..... .... .... .... .... .... ..... .. 22 References .. .... .... .... ..... .... .... .... .... .... ..... .. 25 vii viii Contents Part II Breath Acquisition Systems 3 A Novel Breath Acquisition System Design .... .... .... ..... .. 31 3.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 31 3.2 Breath Analysis . .... ..... .... .... .... .... .... ..... .. 32 3.3 Description of the System .. .... .... .... .... .... ..... .. 33 3.3.1 Breath Gas Collecting... .... .... .... .... ..... .. 35 3.3.2 Signal Sampling ... .... .... .... .... .... ..... .. 36 3.3.3 Data Analysis..... .... .... .... .... .... ..... .. 40 3.4 Experiments.... .... ..... .... .... .... .... .... ..... .. 42 3.4.1 Evaluating Outcomes of Hemodialysis .. .... ..... .. 43 3.4.2 Distinguishing Between Subject Breath Samples ... .. 44 3.5 Results and Discussion..... .... .... .... .... .... ..... .. 46 3.5.1 Results Evaluating Outcomes of Hemodialysis..... .. 46 3.5.2 Results Distinguishing Between Subject Breath Samples .... ..... .... .... .... .... .... ..... .. 47 3.6 Summary.. .... .... ..... .... .... .... .... .... ..... .. 49 References .. .... .... .... ..... .... .... .... .... .... ..... .. 50 4 An LDA-Based Sensor Selection Approach .... .... .... ..... .. 53 4.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 53 4.2 LDA-Based Approach: Definition and Algorithm .... ..... .. 56 4.2.1 Data Expression ... .... .... .... .... .... ..... .. 56 4.2.2 Find Out the Optimum Direction by LDA ... ..... .. 57 4.2.3 Difference Between Two Classes as the Linear Combination of Sensors . .... .... .... .... ..... .. 58 4.2.4 Weight of Sensor .. .... .... .... .... .... ..... .. 60 4.2.5 Algorithm Conclusion... .... .... .... .... ..... .. 60 4.3 Sensor Selection in Breath Analysis System .... .... ..... .. 61 4.3.1 Sensor Selection for Disease Diagnosis . .... ..... .. 61 4.3.2 Evaluating the Medical Treatment . .... .... ..... .. 67 4.4 Comparison Experiment and Performance Analysis... ..... .. 70 4.4.1 Sensor Selection for Disease Diagnosis . .... ..... .. 70 4.4.2 Evaluating the Medical Treatment . .... .... ..... .. 72 4.5 Summary.. .... .... ..... .... .... .... .... .... ..... .. 72 References .. .... .... .... ..... .... .... .... .... .... ..... .. 73 5 Sensor Evaluation in a Breath Acquisition System .. .... ..... .. 77 5.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 77 5.2 System Description .. ..... .... .... .... .... .... ..... .. 78 5.2.1 Framework of the Device.... .... .... .... ..... .. 79 5.2.2 Sensor Array. ..... .... .... .... .... .... ..... .. 79 5.2.3 Sampling Procedure .... .... .... .... .... ..... .. 80 5.2.4 Data Analysis..... .... .... .... .... .... ..... .. 81 5.3 Sensor Evaluation Methods . .... .... .... .... .... ..... .. 81 5.3.1 Cumulative Sensor Importance.... .... .... ..... .. 81 Contents ix 5.3.2 Average Accuracy Improvement... .... .... ..... .. 82 5.3.3 Sensor Inter-relationship. .... .... .... .... ..... .. 83 5.4 Experiments and Discussion. .... .... .... .... .... ..... .. 84 5.4.1 Experiment Configuration.... .... .... .... ..... .. 84 5.4.2 Sensor Evaluation Results.... .... .... .... ..... .. 84 5.4.3 Discussion... ..... .... .... .... .... .... ..... .. 86 5.5 Summary.. .... .... ..... .... .... .... .... .... ..... .. 87 References .. .... .... .... ..... .... .... .... .... .... ..... .. 88 Part III Breath Signal Pre-processing 6 Improving the Transfer Ability of Prediction Models .... ..... .. 91 6.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 91 6.2 Design of Methods... ..... .... .... .... .... .... ..... .. 93 6.2.1 Windowed Piecewise Direct Standardization (WPDS) .... ..... .... .... .... .... .... ..... .. 94 6.2.2 Standardization-Error-Based Model Improvement (SEMI) . .... ..... .... .... .... .... .... ..... .. 96 6.3 Experimental Details . ..... .... .... .... .... .... ..... .. 99 6.3.1 E-nose Module.... .... .... .... .... .... ..... .. 99 6.3.2 Dataset . .... ..... .... .... .... .... .... ..... .. 100 6.3.3 Preprocessing and Feature Extraction... .... ..... .. 102 6.3.4 Data Analysis Procedure. .... .... .... .... ..... .. 102 6.4 Results and Discussion..... .... .... .... .... .... ..... .. 104 6.4.1 Standardization.... .... .... .... .... .... ..... .. 104 6.4.2 Prediction... ..... .... .... .... .... .... ..... .. 106 6.5 Summary.. .... .... ..... .... .... .... .... .... ..... .. 110 References .. .... .... .... ..... .... .... .... .... .... ..... .. 111 7 Learning Classification and Regression Models Based on Transfer Samples.. .... ..... .... .... .... .... .... ..... .. 113 7.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 113 7.2 Related Work... .... ..... .... .... .... .... .... ..... .. 115 7.3 Transfer-Sample-Based Multitask Learning (TMTL) .. ..... .. 116 7.3.1 Transfer-Sample-Based Coupled Task Learning (TCTL). .... ..... .... .... .... .... .... ..... .. 116 7.3.2 TMTL-Parallel and TMTL-Serial .. .... .... ..... .. 120 7.3.3 TMTL-General and the Dynamic Model Strategy... .. 121 7.4 Selection of Transfer Samples ... .... .... .... .... ..... .. 122 7.5 Experiments.... .... ..... .... .... .... .... .... ..... .. 123 7.5.1 Gas Sensor Array Drift Dataset.... .... .... ..... .. 124 7.5.2 Breath Analysis Dataset . .... .... .... .... ..... .. 129 7.5.3 Corn Dataset. ..... .... .... .... .... .... ..... .. 132 7.6 Summary.. .... .... ..... .... .... .... .... .... ..... .. 134 References .. .... .... .... ..... .... .... .... .... .... ..... .. 134 x Contents 8 A Transfer Learning Approach for Correcting Instrumental Variation and Time-Varying Drift.... .... .... .... .... ..... .. 137 8.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 137 8.2 Related Work... .... ..... .... .... .... .... .... ..... .. 138 8.2.1 Autoencoder. ..... .... .... .... .... .... ..... .. 138 8.2.2 Transfer Learning with Autoencoders... .... ..... .. 139 8.3 Drift Correction Autoencoder (DCAE). .... .... .... ..... .. 140 8.3.1 Domain Features... .... .... .... .... .... ..... .. 140 8.3.2 Basic Framework .. .... .... .... .... .... ..... .. 141 8.3.3 Handling Complex Time-Varying Drift . .... ..... .. 142 8.3.4 Summary ... ..... .... .... .... .... .... ..... .. 143 8.4 Experiments.... .... ..... .... .... .... .... .... ..... .. 145 8.4.1 Gas Sensor Array Drift Dataset.... .... .... ..... .. 145 8.4.2 Breath Analysis Dataset . .... .... .... .... ..... .. 150 8.4.3 Corn Dataset. ..... .... .... .... .... .... ..... .. 153 8.4.4 Impact of Different Training Procedures. .... ..... .. 154 8.5 Summary.. .... .... ..... .... .... .... .... .... ..... .. 154 References .. .... .... .... ..... .... .... .... .... .... ..... .. 155 9 Drift Correction Using Maximum Independence Domain Adaptation.. .... .... .... ..... .... .... .... .... .... ..... .. 157 9.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 157 9.2 Related Work... .... ..... .... .... .... .... .... ..... .. 159 9.2.1 Unsupervised Domain Adaptation.. .... .... ..... .. 159 9.2.2 Hilbert–Schmidt Independence Criterion (HSIC) ... .. 160 9.3 Proposed Method.... ..... .... .... .... .... .... ..... .. 160 9.3.1 Domain Feature ... .... .... .... .... .... ..... .. 160 9.3.2 Feature Augmentation... .... .... .... .... ..... .. 161 9.3.3 Maximum Independence Domain Adaptation (MIDA). .... ..... .... .... .... .... .... ..... .. 162 9.3.4 Semi-supervised MIDA (SMIDA).. .... .... ..... .. 164 9.4 Experiments.... .... ..... .... .... .... .... .... ..... .. 165 9.4.1 Synthetic Dataset .. .... .... .... .... .... ..... .. 165 9.4.2 Gas Sensor Array Drift Dataset.... .... .... ..... .. 168 9.4.3 Breath Analysis Dataset . .... .... .... .... ..... .. 170 9.4.4 Corn Dataset. ..... .... .... .... .... .... ..... .. 173 9.5 Summary.. .... .... ..... .... .... .... .... .... ..... .. 175 References .. .... .... .... ..... .... .... .... .... .... ..... .. 177 Part IV Feature Extraction and Classification 10 Feature Selection and Analysis on Correlated Breath Data..... .. 181 10.1 Introduction .... .... ..... .... .... .... .... .... ..... .. 181 10.2 SVM-RFE . .... .... ..... .... .... .... .... .... ..... .. 183 10.2.1 Linear SVM-RFE.. .... .... .... .... .... ..... .. 183 10.2.2 Nonlinear SVM-RFE ... .... .... .... .... ..... .. 184

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.