Deep Learning for Chest Radiographs: Computer-Aided Classification (Primers in Biomedical Imaging Devices and Systems)

✍ Scribed by Yashvi Chandola, Jitendra Virmani, H.S Bhadauria, Papendra Kumar

• Publisher: Academic Press
• Year: 2021
• Tongue: English
• Leaves: 230
• Edition: 1
• Category: Library

✦ Synopsis

Deep Learning for Chest Radiographs enumerates different strategies implemented by the authors for designing an efficient convolution neural network-based computer-aided classification (CAC) system for binary classification of chest radiographs into "Normal" and "Pneumonia." Pneumonia is an infectious disease mostly caused by a bacteria or a virus. The prime targets of this infectious disease are children below the age of 5 and adults above the age of 65, mostly due to their poor immunity and lower rates of recovery. Globally, pneumonia has prevalent footprints and kills more children as compared to any other immunity-based disease, causing up to 15% of child deaths per year, especially in developing countries. Out of all the available imaging modalities, such as computed tomography, radiography or X-ray, magnetic resonance imaging, ultrasound, and so on, chest radiographs are most widely used for differential diagnosis between Normal and Pneumonia. In the CAC system designs implemented in this book, a total of 200 chest radiograph images consisting of 100 Normal images and 100 Pneumonia images have been used. These chest radiographs are augmented using geometric transformations, such as rotation, translation, and flipping, to increase the size of the dataset for efficient training of the Convolutional Neural Networks (CNNs). A total of 12 experiments were conducted for the binary classification of chest radiographs into Normal and Pneumonia. It also includes in-depth implementation strategies of exhaustive experimentation carried out using transfer learning-based approaches with decision fusion, deep feature extraction, feature selection, feature dimensionality reduction, and machine learning-based classifiers for implementation of end-to-end CNN-based CAC system designs, lightweight CNN-based CAC system designs, and hybrid CAC system designs for chest radiographs.

This book is a valuable resource for academicians, researchers, clinicians, postgraduate and graduate students in medical imaging, CAC, computer-aided diagnosis, computer science and engineering, electrical and electronics engineering, biomedical engineering, bioinformatics, bioengineering, and professionals from the IT industry.

✦ Table of Contents

Front Cover
Deep Learning for Chest Radiographs: Computer-Aided Classification
Copyright
Contents
Preface
Acknowledgments
Chapter 1 Introduction
1.1 Motivation
1.2 Introduction to deep learning
1.3 Why deep learning in medical image analysis?
1.4 Medical imaging
1.4.1 Features in medical images
1.4.1.1 Texture features
1.4.1.2 Shape features
1.4.1.3 Color features
1.4.2 Types of medical imaging modalities for analysis of chest tissue
1.4.2.1 Ionizing nature
X-ray
CT scan
PET scan
1.4.2.2 Nonionizing nature
Ultrasound
Magnetic resonance imaging
1.4.3 Why chest radiographs?
1.5 Description of normal and pneumonia chest radiographs
1.6 Objective of the book
1.7 Book chapter outline
References
Further reading
Chapter 2 Review of related work
2.1 Introduction
2.2 Overview of the studies based on the classification of chest radiographs
2.2.1 Overview of machine learning-based studies for the classification of chest radiographs
2.2.1.1 Binary classification-based studies of chest radiographs
2.2.2 Overview of deep learning-based studies for the classification of chest radiographs
2.2.2.1 Binary classification-based studies of chest radiographs
End-to-end pretrained CNN-based CAC system designs for chest radiographs
Hybrid CAC system designs for chest radiographs
Self-designed CNN-based CAC system designs for chest radiographs
2.2.2.2 Multiclass classification-based studies for chest radiographs
End-to-end pretrained CNN-based CAC system designs for chest radiographs
Hybrid CAC system designs for chest radiographs
Self-designed CNN-based CAC system designs for chest radiographs
2.2.2.3 On the basis of chest radiograph dataset used
2.3 Concluding remarks
References
Chapter 3 Methodology adopted for designing of computer-aided classification systems for chest radiographs
3.1 Introduction
3.2 What is a CAC system?
3.3 Need for CAC systems
3.4 Need for CAC systems for chest radiographs
3.5 Types of classifier designs for CAC systems
3.5.1 On the basis of number output classes
3.5.2 On the basis of learning approach
3.6 Deep learning-based CAC system design
3.6.1 On the basis of network connection
3.6.2 On the basis of network architecture
3.7 Workflow adopted in the present work
3.8 Implementation details
3.8.1 Hardware and software specifications
3.8.2 MATLAB Deep Learning Toolbox
3.8.3 Installing Pre-trained networks
3.8.4 Key hyperparameters of deep learning-based networks
3.8.4.1 The problem of overfitting
3.8.4.2 Overcoming the problem of overfitting: Regularization
3.8.5 Key hyperparameters of deep learning-based convolution neural networks used in the present work
3.9 Dataset: Kaggle chest X-ray dataset
3.10 Dataset description
3.11 Dataset generation
3.11.1 Preprocessing module: Image resizing
3.11.2 Dataset bifurcation
3.11.3 Augmentation module: Dataset augmentation
3.11.3.1 Motivation for data augmentation
3.11.3.2 Why augment only the training dataset?
3.11.3.3 How to choose data augmentation technique for medical images
3.11.3.4 Types of data augmentation techniques
3.12 Concluding remarks
References
Further reading
Chapter 4 End-to-end pre-trained CNN-based computer-aided classification system design for chest radiographs
4.1 Introduction
4.2 Experimental workflow
4.3 Transfer learning-based convolutional neural network design
4.4 Architecture of end-to-end pre-trained CNNs used in the present work
4.4.1 Series end-to-end pre-trained CNN model: AlexNet
4.4.2 Directed acyclic graph end-to-end pre-trained CNN model: ResNet18
4.4.3 DAG end-to-end pre-trained CNN model: GoogLeNet
4.5 Decision fusion
4.6 Experiments and results
4.7 Concluding remarks
References
Further reading
Chapter 5 Hybrid computer-aided classification system design using end-to-end CNN-based deep feature extraction and ANFC-LH classifie ...
5.1 Introduction
5.2 Experimental workflow
5.3 Deep feature extraction
5.3.1 GoogLeNet as a deep feature extractor
5.4 Feature selection
5.4.1 Correlation-based feature selection
5.4.2 Feature selection using ANFC-LH
5.5 Adaptive neuro-fuzzy classifier
5.6 Experiment and result
5.7 Concluding remarks
References
Chapter 6 Hybrid computer-aided classification system design using end-to-end Pre-trained CNN-based deep feature extraction and PCA-S ...
6.1 Introduction
6.2 Experimental workflow
6.3 Deep feature extraction
6.4 Feature selection and dimensionality reduction
6.4.1 Correlation-based feature selection
6.4.2 PCA-based feature dimensionality reduction
6.5 SVM classifier
6.6 Experiment and result
6.7 Concluding remarks
References
Further reading
Chapter 7 Lightweight end-to-end Pre-trained CNN-based computer-aided classification system design for chest radiographs
7.1 Introduction
7.2 Experimental workflow
7.3 Lightweight CNN model
7.4 Architecture of lightweight Pre-trained CNN networks used in the present work
7.4.1 DAG lightweight end-to-end Pre-trained CNN model: SqueezeNet
7.4.2 DAG lightweight end-to-end Pre-trained CNN model: ShuffleNet
7.4.3 DAG lightweight end-to-end Pre-trained CNN model: MobileNetV2
7.5 Decision fusion
7.6 Experiments and results
7.7 Concluding remarks
References
Further reading
Chapter 8 Hybrid computer-aided classification system design using lightweight end-to-end Pre-trained CNN-based deep feature extract ...
8.1 Introduction
8.2 Experimental workflow
8.3 Deep feature extraction
8.3.1 Lightweight MobileNetV2 CNN model as deep feature extractor
8.4 Feature selection
8.4.1 Correlation-based feature selection
8.4.2 Feature selection using ANFC-LH
8.5 Adaptive neuro-fuzzy classifier
8.6 Experiment and results
8.7 Concluding remarks
References
Further reading
Chapter 9 Hybrid computer-aided classification system design using lightweight end-to-end Pre-trained CNN-based deep feature extract ...
9.1 Introduction
9.2 Experimental workflow
9.3 Deep feature extraction
9.4 Feature selection and dimensionality reduction
9.4.1 Correlation-based feature selection
9.4.2 PCA-based feature dimensionality reduction
9.5 SVM classifier
9.6 Experiment and results
9.7 Concluding remarks
References
Further reading
Chapter 10 Comparative analysis of computer-aided classification systems designed for chest radiographs: Conclusion and future scope
10.1 Introduction
10.2 Conclusion: End-to-end pretrained CNN-based CAC system design for chest radiographs
10.3 Conclusion: Hybrid CAC system design using end-to-end pretrained CNN-based deep feature extraction and ANFC-LH, PCA- ...
10.4 Conclusion: Lightweight end-to-end pretrained CNN-based CAC system design for chest radiographs
10.5 Conclusion: Hybrid CAC system design using lightweight end-to-end pretrained CNN-based deep feature extraction and A ...
10.6 Comparison of the different CNN-based CAC systems designed in the present work for the binary classification of ches ...
10.7 Future scope
Index
Back Cover