Machine Learning in Medical Imaging and Computer Vision

✍ Scribed by Amita Nandal, Liang Zhou, Arvind Dhaka, Todor Ganchev, Farid Nait-Abdesselam

• Year: 2024
• Tongue: English
• Leaves: 382
• Category: Library

✦ Synopsis

Medical images can highlight differences between healthy tissue and unhealthy tissue and these images can then be assessed by a healthcare professional to identify the stage and spread of a disease so a treatment path can be established. With machine learning techniques becoming more prevalent in healthcare, algorithms can be trained to identify healthy or unhealthy tissues and quickly differentiate between the two. Statistical models can be used to process numerous images of the same type in a fraction of the time it would take a human to assess the same quantity, saving time and money in aiding practitioners in their assessment.

This edited book discusses feature extraction processes, reviews deep learning methods for medical segmentation tasks, outlines optimisation algorithms and regularisation techniques, illustrates image classification and retrieval systems, and highlights text recognition tools, game theory, and the detection of misinformation for improving healthcare provision.

Machine Learning in Medical Imaging and Computer Vision provides state of the art research on the integration of new and emerging technologies for the medical imaging processing and analysis fields. This book outlines future directions for increasing the efficiency of conventional imaging models to achieve better performance in diagnoses as well as in the characterization of complex pathological conditions.

The book is aimed at a readership of researchers and scientists in both academia and industry in computer science and engineering, machine learning, image processing, and healthcare technologies and those in related fields.

✦ Table of Contents

Cover
Contents
About the editors
Preface
1 Machine learning algorithms and applications in medical imaging processing
1.1 Introduction
1.2 Basic concepts
1.2.1 Machine learning
1.2.2 Stages for conducting machine learning
1.2.3 Types of machine learning
1.3 Proposed algorithm for supervised learning based on neuro-fuzzy system
1.3.1 Input factors
1.3.2 Output factors
1.4 Application in medical images (numerical interpretation)
1.5 Comparison of proposed approach with the existing approaches
1.6 Conclusion
References
2 Review of deep learning methods for medical segmentation tasks in brain tumors
2.1 Introduction
2.2 Brain segmentation dataset
2.2.1 BraTS2012-2021
2.2.2 MSD
2.2.3 TCIA
2.3 Brain tumor regional segmentation methods
2.3.1 Fully supervised brain tumor segmentation
2.3.2 Non-fully supervised brain tumor segmentation
2.3.3 Summary
2.4 Small sample size problems
2.4.1 Class imbalance
2.4.2 Data lack
2.4.3 Missing modalities
2.4.4 Summary
2.5 Model interpretability
2.6 Conclusion and outlook
References
3 Optimization algorithms and regularization techniques using deep learning
3.1 Introduction
3.2 Deep learning approaches
3.2.1 Deep supervised learning
3.2.2 Deep semi-supervised learning
3.2.3 Deep unsupervised learning
3.2.4 Deep reinforcement learning
3.3 Deep neural network
3.3.1 Recursive neural network
3.3.2 Recurrent neural network
3.3.3 Convolutional neural network
3.4 Optimization algorithms
3.4.1 Gradient descent
3.4.2 Stochastic gradient descent
3.4.3 Mini-batch-stochastic gradient descent
3.4.4 Momentum
3.4.5 Nesterov momentum
3.4.6 Adapted gradient (AdaGrad)
3.4.7 Adapted delta (AdaDelta)
3.4.8 Root mean square propagation
3.4.9 Adaptive moment estimation (Adam)
3.4.10 Nesterov-accelerated adaptive moment (Nadam)
3.4.11 AdaBelief
3.5 Regularizations techniques
3.5.1 l2 Regularization
3.5.2 l1 Regularization
3.5.3 Entropy regularization
3.5.4 Dropout technique
3.6 Review of literature
3.7 Deep learning-based neuro fuzzy system and its applicability in self-driven cars in hill stations
3.8 Conclusion
References
4 Computer-aided diagnosis in maritime healthcare: review of spinal hernia
4.1 Introduction
4.2 Similar studies and common diseases of the seafarers
4.3 Background
4.4 Computer-aided diagnosis of spinal hernia
4.5 Conclusion
References
5 Diabetic retinopathy detection using AI
5.1 Introduction
5.2 Methodology
5.2.1 Preprocessing
5.2.2 Feature extraction
5.2.3 Classification
5.2.4 Proposed method algorithm
5.2.5 Training and testing
5.2.6 Novel ISVM-RBF
5.3 Results and discussion
5.3.1 Dataset
5.3.2 Image processing results
5.3.3 Comparison with the state-of-the-art studies
5.4 Conclusion
Funding
References
6 A survey image classification using convolutional neural network in deep learning
6.1 Introduction
6.2 Deep learning
6.2.1 Artificial neural network
6.2.2 Recurrent neural network
6.2.3 Feed forward neural network
6.3 Convolutional neural network
6.3.1 Convolutional layer
6.3.2 Pooling layer
6.3.3 Fully connected layer
6.3.4 Dropout layer
6.3.5 Softmax layer
6.4 CNN models
6.4.1 VGGnet
6.4.2 AlexNet
6.4.3 GoogleNet
6.4.4 DenseNet
6.4.5 MobileNet
6.4.6 ResNet
6.4.7 NasNet
6.4.8 ImageNet
6.5 Image classification
6.6 Literature survey
6.7 Discussion
6.8 Conclusion
References
7 Text recognition using CRNN models based on temporal classification and interpolation methods
7.1 Introduction
7.2 Related works
7.3 Datasets
7.4 Model and evaluation matrix
7.4.1 Process of data pre-processing
7.4.2 Air-writing recognition (writing in air)
7.5 Description and working of the model
7.5.1 Handwritten text recognition
7.6 Convolutional neural network
7.7 Connectionist temporal classification
7.8 Decoding
7.9 Optimal fixed length
7.10 Using different interpolation techniques for finding the ideal fixed frame length signals
7.11 CNN architecture
7.12 Evaluation matrix
7.12.1 Handwritten text recognition
7.12.2 Air-writing recognition
7.13 Results and discussion
7.13.1 Handwritten text recognition
7.13.2 Air-writing recognition
7.14 Conclusion
References
8 Microscopic Plasmodium classification (MPC) using robust deep learning strategies for malaria detection
8.1 Introduction
8.1.1 Classification of
using CNN
8.2 Related works
8.3 Methodology
8.3.1 Data preprocessing
8.3.2 Data augmentation
8.3.3 Weight regularization using batch normalization
8.3.4 Classification based on pattern recognition
8.3.5 Models for multi-class classification
8.4 Experimental results and discussion
8.4.1 Dataset description
8.4.2 Performance measures
8.5 Conclusion and future work
References
9 Medical image classification and retrieval using deep learning
9.1 Medical images
9.1.1 Ultrasound images
9.1.2 Magnetic resonance imaging
9.1.3 X-ray imaging for pediatric
9.1.4 X-ray imaging for medical
9.2 Deep learning
9.2.1 Feed-forward neural networks
9.2.2 Recurrent neural networks
9.2.3 Convolutional neural networks
9.3 Deep learning applications in medical images
9.3.1 Identification of anatomical structures
9.3.2 Deep-learning-based organs and cell identification
9.3.3 Deep learning for cell detection
9.4 Deep learning for segmentation
9.5 Conclusion
References
10 Game theory, optimization algorithms and regularization techniques using deep learning in medical imaging
10.1 Introduction
10.2 Game theoretical aspects in MI
10.2.1 Cooperative games
10.2.2 Competitive games
10.2.3 Zero-sum and non-zero-sum games
10.2.4 Deep learning in game theory
10.3 Optimization techniques in MI
10.3.1 Linear programming
10.3.2 Nonlinear programming
10.3.3 Dynamical programming
10.3.4 Particle swarm optimization
10.3.5 Simulated annealing algorithm
10.3.6 Genetic algorithm
10.4 Regularization techniques in MI
10.5 Remarks and future directions
10.6 Conclusion
References
11 Data preparation for artificial intelligence in
federated learning: the influence of artifacts on
the composition of the mammography database
11.1 Introduction
11.2 Federate learning
11.3 Methodology
11.4 Results
11.4.1 Discussion
11.5 Conclusions
References
12 Spatial cognition by the visually impaired: image
processing with SIFT/BRISK-like detector and
two-keypoint descriptor on Android CameraX
12.1 Introduction
12.1.1 Contribution
12.2 Related work
12.3 Methodology
12.3.1 Problem formulation
12.3.2 Identification of all keypoints on the template image: SIFT-like approach
12.3.3 Identification of two keypoints to design the template image feature descriptor: BRISK-like approach
12.3.4 Fast binary feature matching
12.4 Implementation, results, and discussion
12.4.1 Implementation
12.4.2 Results and discussion
12.5 Conclusions
Abbreviations
Data availability
Conflicts of interest
Acknowledgments
References
13 Feature extraction process through hypergraph learning with the concept of rough set classification
13.1 Introduction
13.2 Rough set theory
13.2.1 Preliminaries
13.3 Rough graph
13.4 Proposed work
13.4.1 Rough hypergraph
13.4.2 Methodology
13.4.3 Experimental results
13.5 Results and discussion
References
14 Machine learning for neurodegenerative disease diagnosis: a focus on amyotrophic lateral sclerosis (ALS)
14.1 Introduction
14.2 Neurodegenerative diseases
14.2.1 Alzheimer’s disease
14.2.2 Parkinson’s disease
14.2.3 Huntington’s disease
14.2.4 Amyotrophic lateral sclerosis
14.3 The development stages of NDDs
14.4 Neuroimages on neurodegenerative diseases
14.4.1 Structural magnetic resonance
14.4.2 Diffusion tensor imaging
14.4.3 Functional magnetic resonance imaging
14.5 Machine learning and deep learning applications on ALS
14.6 Proposed research methodology
14.6.1 Methodology flow
14.6.2 Approaches to predictive machine learning
14.6.3 Discussion on review findings
14.7 Conclusion and future work
References
15 Using deep/machine learning to identify patterns and detecting misinformation for pandemics in the post-COVID-19 era
15.1 Introduction
15.2 Literature review
15.2.1 Difference between misinformation and disinformation
15.2.2 Detection of fake news
15.3 Proposed approach
15.3.1 Neural networks
15.3.2 Convolutional neural network
15.3.3 Recurrent neural network
15.3.4 Random forest
15.3.5 Hybrid CNN-RNN-RF model
15.4 Methodology
15.4.1 Datasets
15.4.2 Data-cleaning
15.4.3 Feature extraction method
15.5 Proposed method
15.6 Comparison of models
15.6.1 Hyperparameter optimization method
15.6.2 Evaluation benchmarks
15.7 Future work
15.8 Conclusion
References
16 Integrating medical imaging using analytic modules and applications
16.1 Introduction
16.2 Applications of medical imaging
16.2.1 Radiology and diagnostic imaging
16.2.2 Pathology
16.2.3 Cardiology
16.2.4 Neuroimaging
16.2.5 Ophthalmology imaging
16.3 Key aspects of integrating medical imaging
16.3.1 Interoperability
16.3.2 Picture archiving and communication system
16.3.3 Electronic health records
16.3.4 Decision support systems
16.3.5 Telemedicine and remote access
16.3.6 Clinical workflow optimization
16.4 Analytic modules in integrating medical imaging
16.4.1 Image segmentation
16.4.2 Registration and fusion
16.4.3 Quantitative analysis
16.4.4 Texture analysis
16.4.5 Deep learning and AI algorithms
16.4.6 Visualization and 3D reconstruction
16.5 Algorithm for integrating medical imaging using analytic modules
16.5.1 Data preprocessing algorithms
16.5.2 Segmentation algorithms
16.5.3 Feature extraction methods
16.5.4 Machine learning techniques
16.5.5 Fusion algorithms
16.5.6 Decision support algorithms
16.5.7 Visualization algorithms
16.6 Conclusion
References
Index
Back Cover