Deep Learning in Medical Image Analysis: Challenges and Applications (Advances in Experimental Medicine and Biology, 1213)

✍ Scribed by Gobert Lee (editor), Hiroshi Fujita (editor)

• Publisher: Springer
• Year: 2020
• Tongue: English
• Leaves: 184
• Category: Library

✦ Synopsis

This book presents cutting-edge research and applications of deep learning in a broad range of medical imaging scenarios, such as computer-aided diagnosis, image segmentation, tissue recognition and classification, and other areas of medical and healthcare problems. Each of its chapters covers a topic in depth, ranging from medical image synthesis and techniques for muskuloskeletal analysis to diagnostic tools for breast lesions on digital mammograms and glaucoma on retinal fundus images. It also provides an overview of deep learning in medical image analysis and highlights issues and challenges encountered by researchers and clinicians, surveying and discussing practical approaches in general and in the context of specific problems. Academics, clinical and industry researchers, as well as young researchers and graduate students in medical imaging, computer-aided-diagnosis, biomedical engineering and computer vision will find this book a great reference and very useful learning resource.

✦ Table of Contents

Preface
Contents
Part I Overview and Issues
Deep Learning in Medical Image Analysis
Introduction
Deep Learning for Medical Image Analysis and CAD
Challenges in Deep-Learning-Based CAD
Data Collection
Transfer Learning
Data Augmentation
Training, Validation, and Independent Testing
Acceptance Testing, Preclinical Testing, and User Training
Quality Assurance and Performance Monitoring
Interpretability of CAD/AI Recommendations
Summary
Medical Image Synthesis via Deep Learning
Introduction
Deep Learning Models for Medical Image Synthesis
Convolutional Neural Networks
Generative Adversarial Networks
Within-Modality Synthesis
3D cGAN
Framework
Experimental Results
Locality Adaptive Multi-Modality GANs
Framework
Experimental Results
Cross-Modality Synthesis
3D cGAN with Subject-Specific Local Adaptive Fusion
Framework
Experimental Results
Edge-Aware GANs
Framework
Experimental Results
Conclusion
Part II Applications: Screening and Diagnosis
Deep Learning for Pulmonary Image Analysis: Classification, Detection, and Segmentation
Background of Lung Diseases
Introduction
Methods
Classification of Lung Abnormalities
Detection of Lung Abnormalities
Segmentation of Lung Abnormalities
Conclusion
Deep Learning Computer-Aided Diagnosis for Breast Lesion in Digital Mammogram
Introduction
Related Work
Materials and Methods
Dataset
Datasets Preparation: Training, Validation, and Testing
Preprocessing
Data Balancing and Augmentation
Initialization of Trainable Parameters for Deep Learning Models
Breast Lesion Detection via YOLO
Breast Lesion Segmentation via FrCN
Breast Lesion Classification via Three Convolutional Neural Networks
Experimental Settings
Detection Experimental Settings
Segmentation Experimental Settings
Classification Experimental Settings
Implementation Environment
Experimental Results and Discussion
Evaluation Metrics
Breast Lesion Detection Results
Breast Lesion Segmentation Results
Breast Lesion Classification Results
Conclusion
Decision Support System for Lung Cancer Using PET/CT and Microscopic Images
Introduction
Outline of Decision Support System
Automated Detection of Lung Nodules in PET/CT Images Using Convolutional Neural Network and Radiomic Features
Background
Method Overview
Initial Nodule Detection
False Positive Reduction
Classification Using a Convolutional Neural Network
Handcrafted Radiomic Features
Classification
Results
Image Datasets
Evaluation Metrics
Detection Results
Discussion
Automated Malignancy Analysis of Lung Nodules in PET/CT Images Using Radiomic Features
Introduction
Materials and Methods
Image Dataset
Methods Overview
Volume of Interest (VOI) Extraction
Extraction of Characteristic Features
Classification
Results
Discussion
Automated Malignancy Analysis Using Lung Cytological Images
Introduction
Materials and Methods
Image Dataset
Network Architecture
Results and Discussion
Automated Classification of Lung Cancer Types from Cytological Images
Introduction
Materials and Methods
Image Dataset
Network Architecture
Results and Discussion
Conclusion
Lesion Image Synthesis Using DCGANs for Metastatic Liver Cancer Detection
Introduction
Proposed Method
Dataset
Lesion Image Generation
Method 1: Synthesis Using Poisson Blending
Method 2: Generation Based on a CT Value Distribution
Method 3: Generation Using DCGANs
Selection of the Region of Interest for Lesion Synthesis
Detection Method
Experiments
Results
Discussion
Conclusion
Retinopathy Analysis Based on Deep ConvolutionNeural Network
Introduction
General Arteriolar Narrowing Detection
Blood Vessel Extraction
Related Works
Database
Preprocessing
Blood Vessel Extraction Using DCNN
Detection of Arteriolar Narrowing Using AVR
Related Works
Database
Classification of Arteries and Veins
AVR Measurement
Microaneurysm Detection
Related Work
Database
Methods
Preprocessing
Microaneurysm Detection Based on DCNN
Reducing the Number of False Positives
Examination
Conclusion
Diagnosis of Glaucoma on Retinal Fundus Images Using Deep Learning: Detection of Nerve Fiber Layer Defect and Optic Disc Analysis
Introduction
Related Works
NFLD Detection
Background
Proposed Method
Segmentation Network
Detection Network
Combined Method
Dataset
Preprocessing
Evaluation
Results
Discussion
Optic Disc Analysis
Background
Methods
Dataset
Results
Discussion
Summary
Part III Applications: Emerging Opportunities
Automatic Segmentation of Multiple Organs on 3D CT Images by Using Deep Learning Approaches
Introduction
Issue of Deep Learning for CT Image Segmentation
Two Approaches for Multiple Organ Segmentations Using 2D and 3D Deep CNNs on CT Images
Overview
Deep Learning Anatomical Structures on 2D Sectional Images
Deep Learning Local Appearances of Multiple Organs on 3D CT Images
Conventional Image Segmentation Approach
Results
Discussions
Segmentation Performances
Training Protocol and Transfer Learning
Comparison to Conventional Methods
Computational Efficiency
Conclusion
Techniques and Applications in Skin OCT Analysis
Introduction
Skin Layer Segmentation in OCT
Applications: Roughness, ET
Deep Convolutional Networks in Skin Imaging
Deep Learning for Classification of Dermoscopy Images
Deep Learning for Classification of Full Field OCT Images
Classification of Cross-Sectional OCT 2D Scans
Semantic Segmentation in Cross-Sectional OCT Images
Challenges
Conclusions
Deep Learning Technique for Musculoskeletal Analysis
Importance of Musculoskeletal Analysis and Skeletal Muscle Analysis
Musculoskeletal Recognition by Handcrafted Features and Its Limitations
Skeletal Muscle Segmentation Using Deep Learning
Whole-Body Muscle Analysis Using Deep Learning
Fusion of Deep Learning and Handcrafted Features in Skeletal Muscle Modeling
Conclusion
Index