Fred / Filipe / Gamboa Biomedical Engineering Systems and Technologies

International Joint Conference, BIOSTEC 2009, Porto, Portugal, January 14-17, 2009, Revised Selected Papers
1. Auflage 2010
ISBN: 978-3-642-11721-3
Verlag: Springer
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 416 Seiten

Reihe: Computer Science (R0)

This book contains the refereed proceedings with the best papers of the Second International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2009), held in Porto, Portugal, in January 2009. The 29 revised full papers presented together with two invited papers were carefully reviewed and selected out of 380 paper submissions. BIOSTEC was composed of three co-located conferences: BIODEVICES (International Conference on Biomedical Electronics and Devices) focus on aspects related to electronics and mechanical engineering, especially equipments and materials inspired from biological systems and/or addressing biological requirements. Monitoring devices, instrumentation sensors and systems, biorobotics, micro-nanotechnologies and biomaterials are some of the technologies addressed at this conference. BIOSIGNALS (International Conference on Bio-inspired Systems and Signal Processing) is a forum for those studying and using models and techniques inspired from or applied to biological systems. A diversity of signal types can be found in this area, including image, audio and other biological sources of information. The analysis and use of these signals is a multidisciplinary area including signal processing, pattern recognition and computational intelligence techniques, amongst others. HEALTHINF (International Conference on Health Informatics) promotes research and development in the application of information and communication technologies (ICT) to healthcare and medicine in general and to the specialized support to persons with special needs in particular. Databases, networking, graphical interfaces, intelligent decision support systems and specialized programming languages are just a few of the technologies currently used in medical informatics. Mobility and ubiquity in healthcare systems, standardization of technologies and procedures, certification, privacy are some of the issues that medical informatics professionals and the ICT industry in general need to address in order to further promote ICT in healthcare.

Autoren/Hrsg.

Fred, Ana

Filipe, Joaquim

Gamboa, Hugo

Inhaltsverzeichnis

1;Preface;5
2;Organization;7
3;Table of Contents;10
4;Invited Papers;14
4.1;Computational Intelligence and Image Processing Methods for Applications in Skin Cancer Diagnosis;15
4.1.1;Introduction;15
4.1.2;Basic Image Processing;18
4.1.3;Engineering Decompositions of Images;21
4.1.4;How to Build a Good Classifier?;22
4.1.4.1;Model Types Used in Statistical Learning;22
4.1.4.2;Validation and Model Selection;23
4.1.5;Ensemble Methods;23
4.1.5.1;The Bias/Variance Decomposition for Ensembles;23
4.1.6;Model Training and Cross Validation;25
4.1.7;The ENTOOL Toolbox for Statistical Learning;25
4.1.7.1;ENTOOL Software Architecture;25
4.1.7.2;Primary Models Types;26
4.1.7.3;Secondary Models Types;26
4.1.7.4;Experience in Ensembling;26
4.1.7.5;Feature Selection;26
4.1.7.6;Results;28
4.1.8;Concluding Remarks;30
4.1.9;References;31
4.2;Affective Man-Machine Interface: Unveiling Human Emotions through Biosignals;33
4.2.1;Introduction;34
4.2.2;Background;36
4.2.3;Techniques for Classification;39
4.2.3.1;Analysis of Variance (ANOVA);39
4.2.3.2;Principal Component Analysis (PCA);41
4.2.3.3;k-Nearest Neighbors (k-NN);42
4.2.3.4;Support Vector Machine (SVM);43
4.2.3.5;Artificial Neural Networks (ANN);44
4.2.3.6;Leave-One-Out Cross Validation (LOOCV);45
4.2.4;Recording Emotions;46
4.2.4.1;Participants;46
4.2.4.2;Equipment and Materials;46
4.2.4.3;Procedure;47
4.2.5;Preprocessing;48
4.2.5.1;Normalization;48
4.2.5.2;Baseline Matrix;48
4.2.5.3;Feature Selection;48
4.2.6;Classification Results;49
4.2.6.1;k-Nearest Neighbors (k-NN);50
4.2.6.2;Support Vector Machines (SVM);50
4.2.6.3;Artificial Neural Networks (ANN);51
4.2.6.4;Reflection on the Results;52
4.2.7;Discussion;53
4.2.8;Conclusions;55
4.2.9;References;55
5;Part I BIODEVICES;60
5.1;On-Chip Biosensors Based on Microwave Detection for Cell Scale Investigations;61
5.1.1;Introduction;61
5.1.2;Biosensor RF Design: From Few Cells to Single Cell Detection;63
5.1.2.1;Compact RF Band Stop Resonator Biosensor;63
5.1.2.2;Ultra Sensitive Biosensor Based on a RF Filter Design Approach;65
5.1.2.3;Fabrication Process Details;67
5.1.3;Experimental Protocol for in Vitro RF Characterization;68
5.1.4;Results and Discussion;70
5.1.4.1;Stop Band Resonator Sensor Experiments;70
5.1.4.2;Band Pass RF Filter Sensors Measured Capabilities;71
5.1.5;Conclusions;72
5.1.6;References;72
5.2;Improvements of a Brain-Computer Interface Applied to a Robotic Wheelchair;74
5.2.1;Introduction;74
5.2.2;Background;75
5.2.3;Method;76
5.2.3.1;Graz Dataset;77
5.2.3.2;UAH Dataset;77
5.2.3.3;Feature Extraction: PSD;78
5.2.3.4;Feature Extraction: AAR/RLS;78
5.2.3.5;Classifier: SVM;79
5.2.4;Results;80
5.2.5;Conclusions;82
5.2.6;References;83
5.3;Wavelet-Based and Morphological Analysis of the Globa lFlash Multifocal ERG for Open Angle Glaucoma Characterization;84
5.3.1;Introduction;84
5.3.2;Methods;85
5.3.2.1;Obtaining the Signals;85
5.3.2.2;Study of Severe Lesions by Wavelet Analysis;88
5.3.2.3;Study of Slight Lesions by Morphological Analysis;89
5.3.3;Results;90
5.3.4;Conclusions;93
5.3.5;References;94
5.4;Biotin-Streptavidin Sensitive BioFETs and Their Properties;95
5.4.1;Introduction;95
5.4.2;Method;96
5.4.3;Simulation;98
5.4.4;Results;99
5.4.5;Conclusions;103
5.4.6;References;104
5.5;Improving Patient Safety with X-Ray and Anesthesia Machine Ventilator Synchronization: A Medical Device Interoperability Case Study*;106
5.5.1;Introduction;106
5.5.2;Clinical Use Case;107
5.5.3;Problem Statement and Challenges;108
5.5.4;System Description;109
5.5.4.1;Hardware;110
5.5.4.2;Software;111
5.5.4.3;SOAP;111
5.5.4.4;Synchronization Algorithms;112
5.5.4.5;Alarms;114
5.5.5;Modeling and Verification;115
5.5.6;Code Generation and Implementation;117
5.5.7;Conclusions;118
5.5.8;References;119
5.6;A Ceramic Microfluidic Device for Monitoring Complex Biochemical Reactive Systems;120
5.6.1;Introduction;120
5.6.2;Design;121
5.6.3;Device Fabrication;122
5.6.3.1;Biocompatibility Testing;122
5.6.3.2;Processing Procedures;126
5.6.4;Device Characterization;128
5.6.4.1;Mixing Performance;128
5.6.4.2;Real-Time Analysis of a Complex Diffusion-Reaction Network;130
5.6.5;Conclusions;132
5.6.6;References;132
5.7;Knee Angle Estimation Algorithm for Myoelectric Control of Active Transfemoral Prostheses;134
5.7.1;Introduction;134
5.7.2;Methods;136
5.7.2.1;Adaptive Filter Implementation;138
5.7.2.2;Myoelectric Knee Joint Angle Estimation Algorithm;139
5.7.3;Results;142
5.7.4;Conclusions;144
5.7.5;References;144
5.8;Micro Droplet Transfer between Superhydrophobic Surfaces via a High Adhesive Superhydrophobic Surface;146
5.8.1;Introduction;146
5.8.2;Experimental;147
5.8.2.1;Preparation Method;147
5.8.2.2;Physical Measurements;148
5.8.2.3;Droplet Transfer;149
5.8.3;Results and Discussion;149
5.8.3.1;Superhydrophobic Metal Polymer Surface;149
5.8.3.2;Micro Droplet Transfer;151
5.8.4;Conclusions;152
5.8.5;References;152
6;Part II BIOSIGNALS;153
6.1;Study on Biodegradation Process of Polyethylene Glycol with Exponential Glowth of Microbial Population;154
6.1.1;Introduction;154
6.1.2;Model with Time Dependent Degradation Rate;156
6.1.3;Time Factor of Degradation Rate;159
6.1.4;Simulation with Time Dependent Degradation Rate;160
6.1.5;Discussion;164
6.1.6;References;165
6.2;Variable Down-Selection for Brain-Computer Interfaces;167
6.2.1;Introduction;167
6.2.2;Experimental Design;168
6.2.2.1;Synthetic Data;168
6.2.2.2;Electroencephalogram Data;169
6.2.3;Methods;170
6.2.3.1;Cross-Validation;170
6.2.3.2;Across-Group Variance;171
6.2.4;Results;173
6.2.5;Algorithm Comparison;175
6.2.5.1;Methods;175
6.2.5.2;Data;175
6.2.6;Discussion and Conclusions;178
6.2.7;References;179
6.3;Effect of a Simulated Analogue Telephone Channel onthe Performance of a Remote Automatic System for theDetection of Pathologies in Voice:Impact of Linear Distortions on Cepstrum-Based Assessment - Band Limitation, Frequency Response and Additive Noise;182
6.3.1;Introduction;182
6.3.2;MFCC Formulation;184
6.3.3;Telephone Channel Model;185
6.3.3.1;Amplitude Distortion;185
6.3.3.2;Phase Distortion;186
6.3.3.3;Band Limitation;186
6.3.3.4;Additive Noise;187
6.3.4;Simulation Procedure;188
6.3.4.1;Database;188
6.3.4.2;Classifier;188
6.3.4.3;Testing Protocol;189
6.3.5;Results;189
6.3.5.1;Effect of Band Limitation;189
6.3.5.2;Effect of Amplitude Distortion;190
6.3.5.3;Effect of Phase Distortion;191
6.3.5.4;Effect of Noise Distortion;191
6.3.6;Conclusions;193
6.3.7;References;193
6.4;A Biologically-Inspired Visual Saliency Model to Test Different Strategies of Saccade Programming;196
6.4.1;Introduction;196
6.4.2;Description of the Proposed Models;198
6.4.2.1;Retina Filter;199
6.4.2.2;Cortical-Like Filters;199
6.4.2.3;Normalization and Fusion;202
6.4.2.4;Saccade Programming;202
6.4.3;Experimental Evaluation;203
6.4.3.1;Eye Movement Experiment;203
6.4.3.2;Criterion Choice for Evaluation;204
6.4.3.3;Results;205
6.4.4;Discussion;207
6.4.5;References;208
6.5;Transition Detection for Brain Computer Interface Classification;209
6.5.1;Introduction;209
6.5.2;Description of EEG Data;210
6.5.3;Description of the Method;211
6.5.3.1;Transition Detection;212
6.5.3.2;Choosing the Appropriate Classifier, Once the Transition Has Been Detected;213
6.5.3.3;Procedure to Classify Test Instances with a MovingWindow;213
6.5.3.4;Procedure to Classify Test Instances with a GrowingWindow;214
6.5.3.5;Computing the Size of the MovingWindow;214
6.5.4;Results;215
6.5.5;Summary and Conclusions;218
6.5.6;References;218
6.6;Tuning Iris Recognition for Noisy Images;220
6.6.1;Introduction;220
6.6.2;Iris Recognition;221
6.6.2.1;Daugman’s Approach;221
6.6.3;Image Databases;224
6.6.3.1;Study of UBIRIS and CASIA;226
6.6.4;Modifications and Extensions;226
6.6.4.1;Reflection Removal;226
6.6.4.2;Segmentation;229
6.6.5;Experimental Results;230
6.6.5.1;Segmentation Stage;231
6.6.5.2;Recognition Rate;231
6.6.6;Conclusions;232
6.6.7;References;232
6.7;Three-Dimensional Reconstruction of Macroscopic Features in Biological Materials;234
6.7.1;Introduction;234
6.7.2;Methods;235
6.7.2.1;Stereo 3D Reconstruction;235
6.7.2.2;Auxiliary Depth Estimators;236
6.7.3;Bryophyte Structure;237
6.7.4;Bryophyte Canopy Analysis;238
6.7.4.1;Surface Roughness;239
6.7.5;Results and Discussion;240
6.7.6;Conclusions;241
6.7.7;References;242
6.8;Wavelet Transform Analysis of the Power Spectrum of Centre of Pressure Signals to Detect the Critical Point Interval of Postural Control;244
6.8.1;Introduction;244
6.8.2;Methods;245
6.8.2.1;Subjects;245
6.8.2.2;Centre of Pressure Data;245
6.8.2.3;Data Acquisition and Processing;246
6.8.2.4;Experimental Protocol;246
6.8.2.5;Identifying the Critical Point Using Wavelet Transform Analysis;246
6.8.2.6;Statistical Analysis;249
6.8.3;Results;249
6.8.4;Discussion;251
6.8.5;Conclusions;252
6.8.6;References;252
6.9;Early Detection of Severe Apnoea through Voice Analysis and Automatic Speaker Recognition Techniques;254
6.9.1;Introduction;254
6.9.2;Physiological and Acoustic Characteristics in OSA Speakers;256
6.9.2.1;Initial Contrastive Acoustic Study;257
6.9.3;Apnoea Database;258
6.9.3.1;Speech and Image Collection;259
6.9.3.2;Speech Corpus;259
6.9.4;Apnoea Voice Modelling with GMM;260
6.9.4.1;GMM Training and Testing Protocol;261
6.9.4.2;A Study of Apnoea Speaker Resonance Anomalies Using GMMs;262
6.9.5;Automatic Diagnosis of Severe Apnoea Using GMMs;263
6.9.6;Conclusions and Future Research;264
6.9.7;References;265
6.10;Automatic Detection of Atrial Fibrillation for Mobile Devices;267
6.10.1;Introduction;267
6.10.2;Atrial Fibrillation;267
6.10.3;State of the Art;268
6.10.4;Datasets;269
6.10.5;Algorithm;269
6.10.5.1;ECG Premachining;270
6.10.5.2;PPV-Detector;270
6.10.5.3;PPV-MF-Detector;272
6.10.6;Results;276
6.10.7;Conclusions and Discussion;278
6.10.8;Outlook;278
6.10.9;References;279
6.11;Speaker-Adaptive Speech Recognition Based on SurfaceElectromyography;280
6.11.1;Introduction;280
6.11.2;The EMG-PIT Data Corpus;281
6.11.3;EMG-Based Speech Recognizer;283
6.11.3.1;Initial Time Alignment;283
6.11.3.2;Feature Extraction;284
6.11.3.3;Training Process;285
6.11.3.4;Across-Speaker Experiments and Adaptation;285
6.11.3.5;Bundling of Phonetic Features;285
6.11.3.6;Testing;287
6.11.4;Experimental Results;287
6.11.4.1;Preprocessing for the Speaker-Dependent System;287
6.11.4.2;Cross-Speaker and Adaptive Experiments;289
6.11.4.3;Summary;291
6.11.5;Conclusions;292
6.11.6;References;293
6.12;Towards the Development of a Thyroid Ultrasound Biometric Scheme Based on Tissue Echo-morphological Features;295
6.12.1;Introduction;295
6.12.2;Problem Formulation;297
6.12.3;Feature Extraction Module;298
6.12.3.1;Segmentation;298
6.12.3.2;Feature Extraction;298
6.12.3.3;Rayleigh Distribution Parameters from Original Sample and Speckle Field;300
6.12.3.4;Wavelet Energy Coefficients;301
6.12.3.5;Radon Transform Features;301
6.12.3.6;Dimensionality Reduction;302
6.12.4;Classification Module;303
6.12.4.1;K-Nearest Neighbors Classifier;303
6.12.4.2;MAP Classifier;303
6.12.4.3;Minimum Entropy Distance Classifier;304
6.12.5;Results and Discussion;305
6.12.6;Conclusions;306
6.12.7;References;307
7;Part III HEALTHINF;308
7.1;Collecting, Analyzing, and Publishing Massive Data about the Hypertrophic Cardiomyopathy*;309
7.1.1;Introduction;309
7.1.2;Operative Process;311
7.1.3;System Architecture and Technology;313
7.1.3.1;System Architecture;313
7.1.3.2;DataModel;315
7.1.3.3;Technology;317
7.1.4;User Interfaces;317
7.1.5;Use Interest;319
7.1.6;Conclusions and Future Work;320
7.1.7;References;320
7.2;BredeQuery: Coordinate-Based Meta-analytic Search of Neuroscientific Literature from the SPM Environment;322
7.2.1;Introduction;322
7.2.2;Brede Database;324
7.2.3;Related Tools;325
7.2.4;Software Description;326
7.2.5;Example Session;328
7.2.6;Future Work;330
7.2.7;Conclusions;331
7.2.8;References;331
7.3;Simulation of ECG Repolarization Phase with Improved Model of Cell Action Potentials;333
7.3.1;Introduction;333
7.3.2;Methods;335
7.3.2.1;Model of the Left Ventricle;335
7.3.2.2;Simulation Method;335
7.3.2.3;Simulation Results –Wohlfart AP Model;337
7.3.3;ModifiedAPModel;338
7.3.4;Conclusions;339
7.3.5;References;340
7.4;Advances in Computer-Based Autoantibodies Analysis;341
7.4.1;Introduction;341
7.4.2;Background and Motivations;342
7.4.2.1;ANA Tests;342
7.4.2.2;Anti-dsDNA Tests;343
7.4.3;System Architecture;344
7.4.3.1;ANA Image Classification;344
7.4.3.2;Anti-dsDNA Image Classification;347
7.4.4;DataSet;348
7.4.5;Experimental Evaluation;349
7.4.5.1;Results of ANA Classification;349
7.4.5.2;Results of Anti-dsDNA Classification;352
7.4.6;Conclusions;352
7.4.7;References;353
7.5;Support Vector Machine Diagnosis of Acute Abdominal Pain;355
7.5.1;Introduction;355
7.5.2;Methods;356
7.5.2.1;Data Acquisition;356
7.5.2.2;Variable Transformation;357
7.5.2.3;Data Partitioning;357
7.5.2.4;Support Vector Machines;357
7.5.2.5;Variable Ranking;358
7.5.2.6;Class Imbalance Correction;358
7.5.2.7;PerformanceMeasure;358
7.5.3;Results;359
7.5.3.1;Classification Performance;359
7.5.3.2;Variable Selection;360
7.5.4;Discussion;362
7.5.5;Conclusions;362
7.5.6;References;363
7.6;Near Field Communication and Health: Turning a Mobile Phone into an Interactive Multipurpose Assistant in Healthcare Scenarios;364
7.6.1;Introduction;364
7.6.2;NFC Technology;365
7.6.2.1;The Origins: RFID;365
7.6.2.2;NFC Technological Features;366
7.6.3;NFC in the RealWorld;367
7.6.3.1;NFC Devices;367
7.6.3.2;NFC Applications;368
7.6.3.3;NFC Standards and Organizations;369
7.6.4;NFC in Sanitary Environments;370
7.6.4.1;Service Initiation Applications in Sanitary Environments;371
7.6.4.2;Peer-to-Peer Applications in Sanitary Environments;373
7.6.4.3;An End-User Application: The Telephone as an Electronic Case History;374
7.6.5;NFC Devices Used in the Realization of the System;375
7.6.6;Conclusions and Future Work;375
7.6.7;References;376
7.7;Electronic Health Records: An Enhanced Security Paradigm to Preserve Patient’s Privacy;377
7.7.1;Introduction;377
7.7.2;Motivation;378
7.7.2.1;AttackerModels;378
7.7.3;Security and Privacy Objectives for EHR’S;379
7.7.3.1;Anonymity;379
7.7.3.2;Authentication;380
7.7.3.3;Authorization;380
7.7.3.4;Confidentiality;381
7.7.3.5;Deniability;381
7.7.3.6;Unlinkability;381
7.7.3.7;Data Structure;382
7.7.4;Investigation of EHR Systems;382
7.7.4.1;Virtual Hard Disk;382
7.7.4.2;Personal Health Record Platforms;383
7.7.4.3;PIPE;383
7.7.5;Privacy Enhanced EHR;384
7.7.5.1;Anonymous Communication;384
7.7.5.2;Anonymous Authentication;384
7.7.5.3;Authorization and Confidentiality;385
7.7.5.4;Pseudonymization;385
7.7.5.5;Multiple Identities;386
7.7.6;Conclusions;387
7.7.7;References;387
7.8;Augmented Feedback System to Support Physical Therapy of Non-specific Low Back Pain;389
7.8.1;Introduction;389
7.8.2;Background;390
7.8.3;Identification of Sensor Configuration;391
7.8.4;The Backtrainer System;393
7.8.4.1;System Overview;393
7.8.4.2;Inertial Sensors;393
7.8.4.3;Software;393
7.8.4.4;Exercises;395
7.8.5;Clinical Evaluation;398
7.8.6;Conclusions and Future Work;400
7.8.7;References;401
7.9;Multi-analytical Approaches Informing the Risk of Sepsis;402
7.9.1;Introduction;402
7.9.2;Data in Study;403
7.9.2.1;Ethical Review, Funding and Data Ownership;404
7.9.2.2;Structure of Paper;404
7.9.3;Decision Tree Approach;404
7.9.3.1;Entropy;404
7.9.3.2;Information Gain;405
7.9.4;Cluster Analysis;406
7.9.4.1;K-Means Clustering;406
7.9.4.2;Modified K-Means;407
7.9.5;Comparative Analysis;407
7.9.5.1;Regression Analysis;407
7.9.5.2;Decision Tree Analysis;409
7.9.5.3;Cluster Analysis;409
7.9.6;Discussion;410
7.9.7;Conclusions;413
7.9.8;References;413
8;Author Index;415