Handbook of Artificial Intelligence in Healthcare: Vol. 1 – Advances and Applications

Preface
Contents
Part I Advances in AI for Healthcare Signal, Image, and Video Processing
1 Advances in Artificial Intelligence for the Identification of Epileptiform Discharges
	1.1 Background
	1.2 Artificial Intelligence Tools
	1.3 Pre-Ictal, Ictal, Post-Ictal Detection
	1.4 Seizure Detection
	1.5 Inter-Ictal Identification
	1.6 Seizure Onset Zone
	1.7 Implications and Future Challenges
	References
2 Characterizing EEG Electrodes in Directed Functional Brain Networks Using Normalized Transfer Entropy and PageRank
	2.1 Introduction
	2.2 Current Approaches to Study Directional Information Flow in FBNs
		2.2.1 Normalized Transfer Entropy
		2.2.2 PageRank
	2.3 Materials and Methods
		2.3.1 Experimental Design
		2.3.2 EEG Data Acquisition and Pre-Processing
		2.3.3 Behavioral Data
		2.3.4 Directed Information Flow Using Normalized Transfer Entropy
		2.3.5 Rate of Change of Cognition
	2.4 Experimental Results and Discussion
		2.4.1 Electrode Wise Analysis
		2.4.2 Observation Phase
		2.4.3 Entire Population Group-Wise Analysis
	2.5 Conclusion
	References
3 Autistic Verbal Behavior Language Parameterization
	3.1 Introduction
	3.2 Considerations About the Autistic Spectrum Disorder
		3.2.1 Degrees of Autism
		3.2.2 Verbal Behavior
	3.3 Materials and Methods
		3.3.1 Hardware
		3.3.2 Protocol
		3.3.3 Software
	3.4 Preliminary Evaluation
		3.4.1 Modeling the Problem with Metadata
		3.4.2 Advantages and Disadvantages of Using the Proposed Approach
	3.5 The Sounds of the Use-Case
	3.6 Test and Evaluation
		3.6.1 Pre-processing
		3.6.2 Variable Selection
		3.6.3 Automatic Timestamp Detection
		3.6.4 Model for Timestamp Detection
		3.6.5 Model Findings and Results Analysis
	3.7 Conclusions and Future Work
	References
4 Case Studies to Demonstrate Real-World Applications in Ophthalmic Image Analysis
	4.1 Introduction
	4.2 Related Work
		4.2.1 Retinal Image Quality Assessment
		4.2.2 Arteriolar-to-Venular Index and A/V Classification
		4.2.3 Retinopathy of Prematurity
	4.3 Case Study: Retinal Quality Assessment
		4.3.1 Dataset
		4.3.2 Methods
		4.3.3 Results
	4.4 Case Study: Arteriolar-to-Venular Index
		4.4.1 Datasets
		4.4.2 Methods
		4.4.3 Results
	4.5 Case Study: Retinopathy of Prematurity
		4.5.1 Datasets
		4.5.2 Methods
		4.5.3 Results
	4.6 Summary
	References
5 Segmentation of Petri Plate Images for Automatic Reporting of Urine Culture Tests
	5.1 Introduction
	5.2 Related Work
	5.3 Automatic Petri Plate Analysis Pipeline
		5.3.1 Image Acquisition
		5.3.2 Segmentation
		5.3.3 Colony Classification and Count
	5.4 Conclusions
	References
6 Repurposing Routine Imaging for Cancer Biomarker Discovery Using Machine Learning
	6.1 Introduction
		6.1.1 Imaging Modalities in Cancer Care
		6.1.2 Imaging in the Cancer Pathway
	6.2 Cancer Biomarker Research
	6.3 Machine Learning Applications in Cancer Cross-Sectional Imaging
		6.3.1 Lesion Detection/Classification
		6.3.2 Segmentation
		6.3.3 Cancer-Related Radiomics
	6.4 Preparing Radiology Data for Machine Learning
	6.5 Example of Biomarker Discovery: Sarcopenia in Cancer
		6.5.1 Defining Cancer Sarcopenia
		6.5.2 Scalable Solutions to Radiological Sarcopenia Assessment
		6.5.3 Remaining Translational Gaps
	6.6 Conclusion
	References
7 Automatic Detection of LST-Type Polyp by CNN Using Depth Map
	7.1 Introduction
	7.2 Background
		7.2.1 Removal of Specular Reflectance Components and Generation of Lambertian Images
		7.2.2 Recovering 3D Shape and Creating Depth Map
	7.3 Construction of U-Net Using Depth Map
		7.3.1 Preprocessing and Construction of Dataset
		7.3.2 Construction of CNN Model Using U-Net Structure
	7.4 Experiment
		7.4.1 Evaluation Method
		7.4.2 Detection Experiment
	7.5 Conclusion
	References
8 Artificial Intelligence and Deep Learning, Important Tools in Assisting Gastroenterologists
	8.1 Introduction
	8.2 Computer-Assisted Colonoscopy for CRC Early Detection
		8.2.1 Polyps’ Semantic Segmentation
		8.2.2 Reviews and Meta-Analysis, Randomized Studies and AI Embedded Colonoscopy Devices
		8.2.3 Well Structured Labeled Databases
	8.3 Dealing with Video Colonoscopy Frames
	8.4 Deep Learning on Video Colonoscopies
		8.4.1 Deep Learning on Video Colonoscopies Using Nvidia Jetson Xavier
	8.5 Conclusions
	References
9 Last Advances on Automatic Carotid Artery Analysis in Ultrasound Images: Towards Deep Learning
	9.1 Introduction
	9.2 Carotid Artery Segmentation and Intima Media Thickness Estimation in Ultrasound Images
		9.2.1 Deep Learning Proposal for IMT Estimation and Plaque Detection
	9.3 Carotid Artery Plaque Classification and Risk Assessment in 2D CA Ultrasound Images
		9.3.1 Data Properties: Transversal/Follow-Up, Different Devices, Image Modality, Artery Territory, Number of Samples and Ground Truth
		9.3.2 Work Objectives
		9.3.3 Image Features
		9.3.4 Methods and Results
	9.4 Discussion: Challenges in Deep Learning
	9.5 Conclusions and Future Perspective
	9.6 Appendix
	References
10 Radiomics and Its Application in Predicting Microvascular Invasion of Hepatocellular Carcinoma
	10.1 Introduction
		10.1.1 What is Radiomics
		10.1.2 What Has Been Achieved in Medical Image Analysis Using Radiomics
		10.1.3 Application of Radiomics in Hepatocellular Carcinoma
	10.2 Radiomics Signature and Prediction Model
		10.2.1 Medical Image Acquisition
		10.2.2 Calibration and Segmentation of Tumour Regions
		10.2.3 Feature Extraction and Quantification
		10.2.4 Feature Selection
		10.2.5 Classification and Prediction
		10.2.6 Material and Clinical Model
		10.2.7 Radiomics Model and Fusion Model for Predicting MVI
	10.3 Experiment
		10.3.1 Experimental Result
		10.3.2 The Direction of Future Progress
	10.4 Conclusion
	References
11 Artificial Intelligence in Remote Photoplethysmography: Remote Heart Rate Estimation from Video Images
	11.1 Introduction
	11.2 Naive Methods
	11.3 Blind Signal Separation
		11.3.1 Independent Component Analysis
		11.3.2 Principal Component Analysis
		11.3.3 Joint Blind Signal Separation
	11.4 Modelling
		11.4.1 CHROM
		11.4.2 Illumination Rectification
		11.4.3 2SR, POS
		11.4.4 Motion Reduction
	11.5 Deep Learning
		11.5.1 Feature Extraction and Representation
		11.5.2 Interference Separation and Signal Enhancement
	11.6 Popular Datasets for rPPG Learning
	11.7 Future
	References
Part II Advances in AI for Healthcare Information and Data Analytics
12 Mining Data to Deal with Epidemics: Case Studies to Demonstrate Real World AI Applications
	12.1 Introduction
		12.1.1 Goal and Research Questions
		12.1.2 Introduction to Data Mining
		12.1.3 Data Mining Techniques
		12.1.4 Chapter Overview
	12.2 Literature Review
		12.2.1 Dengue Fever Analysis and Prediction with Classification and Association Rules
		12.2.2 Mumps Analysis with Clustering and Association Rules
		12.2.3 Cholera Analysis with Classification and Association Rules
		12.2.4 Measles Analysis with Classification
		12.2.5 Ebola Analysis with Clustering
	12.3 Methodology
		12.3.1 Methodology Outline
	12.4 Experiments
		12.4.1 Dataset
		12.4.2 Classification
		12.4.3 Clustering
		12.4.4 Association Rule Mining
	12.5 Conclusion
		12.5.1 Discussion
		12.5.2 Overview of Contribution
		12.5.3 Future Directions
	References
13 A Powerful Holonic and Multi-Agent-Based Front-End for Medical Diagnostics Systems
	13.1 Introduction
	13.2 State of the Art
	13.3 Differential Diagnosis and the Holonic Medical Diagnostics System (HMDS)
		13.3.1 Differential Diagnosis as a Holonic Domain
		13.3.2 The Holonic Medical Diagnostics System
	13.4 Learning in the HMDS
	13.5 Simulations
		13.5.1 The Assessment of the Diagnosis Abilities
		13.5.2 The Assessment of the Self-Organization Abilities
	13.6 Discussion
	13.7 Conclusion
	References
14 Computer-Aided Detection of Depressive Severity Using Multimodal Behavioral Data
	14.1 Introduction
	14.2 Multimodal Behavioral Dataset of Chinese University Students with and Without Depressive Tendencies
		14.2.1 Collecting Survey Data
		14.2.2 Acquiring Behavioral Data
	14.3 Computer-Aided Detection of Depressive Severity
		14.3.1 Feature Extraction
		14.3.2 Detection Model
	14.4 Performance Evaluation
		14.4.1 Experimental Setup
		14.4.2 Evaluation Functions
		14.4.3 Results
	14.5 Conclusions
	References
15 Classifying Process Traces for Stroke Management Quality Assessment: A Deep Learning Approach
	15.1 Introduction
	15.2 Background
		15.2.1 Convolutional Neural Networks
		15.2.2 Autoencoders
		15.2.3 Recurrent Neural Networks
	15.3 Related Work
	15.4 Deep Learning Process Trace Classification for Quality Assessment
	15.5 Experimental Results
	15.6 Discussion and Conclusions
	References
16 Synergy-Net: Artificial Intelligence at the Service of Oncological Prevention
	16.1 Introduction
	16.2 Synergy-Net
		16.2.1 Medical Imaging and AI
		16.2.2 The Synergy-Net Architecture
		16.2.3 Synergy-Net: Analysed Tumours
	16.3 Skin Cancer
	16.4 Lung
	16.5 Colon Rectum Cancer
	16.6 Breast Cancer
	16.7 Gastric Carcinoma
	16.8 Thyroid Cancer
	16.9 Prostate Cancer
	16.10 Conclusions and Future Perspectives
	References
17 New Insights on Implementing and Evaluating Artificial Intelligence in Cardiovascular Care
	17.1 Introduction
		17.1.1 Artificial Intelligence and Machine Learning
		17.1.2 Relevance of Artificial Intelligence to the Future of Cardiovascular Care Delivery
		17.1.3 Implementing AI Within Institutional Healthcare Environments
	17.2 Data Capture and Management
		17.2.1 Data Availability
		17.2.2 Data Quality
		17.2.3 Data Generalizability
		17.2.4 Missing Data
		17.2.5 Data Permission and Privacy
	17.3 Model Development and Validation
		17.3.1 Model Development
		17.3.2 Model Performance Metrics May not Reflect Clinical Applicability
		17.3.3 Model Generalizability and Explainability
		17.3.4 Algorithmic Bias and Equity, Diversity and Inclusion
	17.4 Clinical Integration and Support
		17.4.1 Human Barriers
		17.4.2 Regulatory Considerations and Demonstrating Patient Value
	17.5 Chapter Summary
	References