A Beginner's Guide to Medical Application Development with Deep Convolutional Neural Networks

Cover
Half Title
Title Page
Copyright Page
Table of Contents
Preface
Acknowledgments
About the Authors
1 Introduction to Medical Data and Image Analysis
	1.1 Introduction
	1.2 The Classical Vs Modern Medical Imaging Technology
		1.2.1 The History of ECG
		1.2.2 The History of X-Ray
		1.2.3 Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) Scan
	1.3 Medical Image Preprocessing
	1.4 Medical Image Analysis
	1.5 Application of the Generative Pre-Trained Transformer (GPT) in Image Analysis
	1.6 How to Use  This Book
	1.7 Conclusion
	References
2 The Convolutional Neural Network
	2.1 Introduction
		2.1.1 The Visual Cortex of the Brain
	2.2 Advancements in the Field of Deep Learning and Convolutional Neural Networks
	2.3 Implementation of a Basic Convolutional Neural Network Using Python and TensorFlow
	2.4 Transfer Learning in Convolutional Neural Networks
		2.4.1 Introduction About Transfer Learning
		2.4.2 Some of the Famous Transfer Learning Examples of Convolutional Neural Network
			2.4.2.1 Inception V3 Aka GoogLeNet
			2.4.2.2 The Residual Network Architecture (Aka ResNet-50)
			2.4.2.3 The VGG16 and VGG19 Architecture
	2.5 A Comparative Study Between Advanced CNN Systems, Generative Adversarial Neural Network (GANN), and Stable Diffusion (SD) Model
	2.6 Conclusion
	References
3 The Detection of COVID-19 Pneumonia Using Inception V3 and Custom Designed Bi-Modal Looping DCNN Via Analysis of X-Ray Images
	3.1 Introduction
	3.2 Related Research
	3.3 Dataset Description and Programming Libraries Overview
		3.3.1 Data-Flow Modeling
		3.3.2 Parameters of the Image Data Extractor Function
		3.3.3 Implementation of the Image Data Extractor Function in Python
	3.4 Methodology
		3.4.1 The Inception V3 Architecture and the Bi-Modal Looping DCNN Architecture
		3.4.2 Implementation of Inception V3 Using Python and Transfer Learning
		3.4.3 The Adaptive Moments Optimization Technique (Aka Adam Optimizer)
			3.4.3.1 The Adaptive Moments Optimization Aka ADAM Optimizer Algorithm
			A. Moving Averages of Gradient and Squared Gradient
			B. Bias Correction for the First Momentum Estimator
			C. Final Formulation of the ADAM Optimizer
	3.5 Result Analysis
	3.6 Conclusion
	References
4 Detection of Pneumonia From a Small-Scale Dataset of X-Ray Images of Lungs By Using a Compound Batch-Normalizing Convolutional Neural Feature Extracting Random Forest Classifier
	4.1 Introduction
	4.2 Pneumonia in Lungs: A Short Guide
		4.2.1 Biomedical Agents of Pneumonia
			4.2.1.1 Pneumonia Due to Bacteria
			4.2.1.2 Pneumonia Due to Viruses
			4.2.1.3 Pneumonia Due to Fungus
		4.2.2 Deep Learning and Its Applications
			4.2.2.1 Deep Transfer Learning
	4.3 The Data Preprocessing Function Creation
	4.4 The VGG16 Architecture
	4.5 Optimization Using the Nesterov’s Adaptive Moments Optimization Algorithm (NADAM Optimizer)
		4.5.1 The Final Formulation of NADAM Optimization Algorithm
	4.6 The Python Coding Overview and Guidelines for VGG16 Transfer Learning Approach
	4.7 The Python Coding Overview and Guidelines for Custom DCNN Approach (From Scratch)
	4.8 Conclusion
	References
5 An Adaptive Profound Transfer Learning Strategy for Malaria Cell Parasite Classification and Detection
	5.1 Introduction
	5.2 Related Research
	5.3 Methodology
		5.3.1 The Convolutional Neural Network Using Keras With TensorFlow Back-End
		5.3.2 The Inception V3 Architecture
		5.3.3 Algorithmic Approach and Working Principle of the ADAM Optimizer
			5.3.3.1 Bias Correction for the First Momentum Estimator
		5.3.4 The Concept of the Ensemble Creation
	5.4 Result Analysis
	5.5 Discussion
	5.6 Conclusion
	References
6 Implementation of a Deep Convolutional Auto-Encoding Image-Reconstruction Network (DCARN) to Visualize Distinct Categories of COVID-19 and Pneumonia X-Ray Image Features
	6.1 Introduction
	6.2 The Auto-Encoder Architecture
		6.2.1 Convolutional Auto-Encoder
		6.2.2 The Mathematical Modeling of a Convolutional Auto-Encoder
		6.2.3 The Convolutional Encoder of the CAEs
		6.2.4 The Convolutional Decoder of the CAEs
	6.3 Implementation of the Proposed DCARN (Aka Deep Convolutional Auto-Encoding Image Reconstruction Network)
	6.4 The Data Preprocessing for the DCARN
	6.5 The DCARN Implementation Using TensorFlow in Python
	6.6 The Result Analysis By Plotting Performance Metrics Graphs
	6.7 Conclusion
	References
7 Super Resolution Generative Adversarial Neural Network (SR-GANN) With Bi-Modal Multi-Perceptron Layers for Medical X-Ray Images
	7.1 Introduction
	7.2 Generative Adversarial Neural Networks
		7.2.1 Mathematical Modeling of Generative Adversarial Neural Network
		7.2.2 Notation for GANN
		7.2.3 The Discriminator
		7.2.4 The Generator
		7.2.5 Binary Cross Entropy
		7.2.6 Minor Caveats
		7.2.7 Model Optimization
		7.2.8 Why Was This Necessary?
	7.3 The Image Super Resolution Generative Adversarial Neural Network for Super Resolution of X-Ray Images of Around 5,000 Patients
		7.3.1 The Generator Can Be Trained in Two Ways
			7.3.1.1 Contrastive Divergence
			7.3.1.2 Adversarial Training
	7.4 Implementation of the X-Ray Image Super Resolution Generative Neural Network System
	7.5 Conclusion
	References
8 Conclusion
Index