This book offers detailed information on biomedical imaging using Deep Convolutional Neural Networks (Deep CNN). It focuses on different types of biomedical images to enable readers to understand the effectiveness and the potential. It includes topics such as disease diagnosis and image processing perspectives., Deep Learning in Biomedical Signal and Medical Imaging discusses classification, segmentation, detection, tracking, and retrieval applications of non-invasive methods such as EEG, ECG, EMG, MRI, fMRI, CT, and X-RAY, among others. It surveys the most recent techniques and approaches in this field, with both broad coverage and enough depth to be of practical use to working professionals. It includes examples of the application of signal and image processing employing Deep CNN to Alzheimer’s disease, brain tumor, skin cancer, breast cancer, and stroke prediction, as well as ECG and EEG signals. This book offers enough fundamental and technical information on these techniques, approaches, and related problems without overcrowding the reader’s head. It presents the results of the latest investigations in the field of Deep CNN for biomedical data analysis. The techniques and approaches presented in this book deal with the most important and/or the newest topics encountered in this field. They combine the fundamental theory of artificial intelligence (AI), machine learning (ML), and Deep CNN with practical applications in biology and medicine. Certainly, the list of topics covered in this book is not exhaustive, but these topics will shed light on the implications of the presented techniques and approaches on other topics in biomedical data analysis.
Author(s): Ngangbam Herojit Singh & Utku Kose & Sarada Prasad Gochhayat
Publisher: CRC Press
Year: 2024
Language: English
Pages: 274
Cover
Half Title
Series
Title
Copyright
Contents
About the Editors
List of Contributors
Chapter 1 Detection of Diabetic Retinopathy from Retinal Fundus Images by Using CNN Model ResNet-50
1.1 Introduction
1.2 Related Works
1.3 CNN Model
1.3.1 History
1.3.2 Basic Architecture
1.3.3 Opportunities, Challenges, and Applications
1.4 The Proposed CNN Model
1.4.1 ResNet-50 Architecture
1.4.2 Implementation Steps of ResNet-50 Model
1.5 Results and Discussion
1.6 Conclusion and Future Direction
References
Chapter 2 DNASNet-RF: Automated Deep NAS-Network with Random Forest for Classifying and Detecting Multi-Class Brain Tumor
2.1 Introduction
2.1.1 Motivation
2.1.2 Problem Statement
2.1.3 Applications of Medical Imaging
2.2 Experimental Design and Implementation
2.2.1 Dataset Description
2.2.2 Environmental Setup
2.2.3 Deep Learning-Based NASNetLarge
2.3 Results and Discussion
2.3.1 Performance Evaluation Matrices
2.3.2 Experimental Results
2.3.3 Boundary Box Detection Using Gaussian Smoothing
2.3.4 Comparison with Current SOTA Models
2.3.5 Discussion
2.4 Conclusion and Future Work
References
Chapter 3 Deep CNNs in Image-Guided Diagnosis of Breast and Skin Cancers
3.1 Introduction
3.2 Fundamentals of Deep CNN
3.2.1 Input Layer
3.2.2 Convolutional Layers
3.2.3 Activation Functions
3.2.4 Pooling Layers
3.2.5 Flattening
3.2.6 Fully Connected (Dense) Layer
3.2.7 Output Layer
3.2.8 Loss Functions
3.2.9 Regularizations
3.2.10 Optimizations
3.3 Breast Cancer Diagnosis Using Deep CNNs
3.3.1 Related Databases
3.3.2 Image Acquisition and Pre-Processing
3.3.3 Image Analysis
3.3.4 Multimodal Approaches and Fusion Techniques
3.4 Skin Cancer Diagnosis with Deep CNNs
3.4.1 Related Databases
3.4.2 Dermoscopy Image Analysis
3.4.3 Related Works
3.4.4 Computer-Aided Diagnosis of Skin Conditions—A Case Study
3.5 Interpretability and Explainability in Deep CNNs
3.5.1 Gradient-Weighted Class Activation Mapping (Grad-CAM)
3.5.2 Saliency Maps and Sensitivity Analysis
3.5.3 LIME and SHAP
3.5.4 Concept Activation Vectors
3.5.5 Integrated Gradients
3.6 Future Scopes
3.7 Conclusion
References
Chapter 4 Robust Learning Principle Design to Detect Diabetic Retinopathy Disease in Early Stages with Skilled Feature Extraction Policy
4.1 Introduction
4.2 Related Works
4.3 Methodology
4.4 Image Preprocessing
4.5 Data Augmentation
4.6 Image Segmentation
4.7 Deep Neural Networks
4.8 Hybrid DGFM
4.9 Firefly Algorithms
4.10 Results and Discussion
4.11 Experimental Results
4.12 Conclusion and Future Work
References
Chapter 5 Liver Tumour Detection Using Machine Learning Techniques: A Systematic Review
5.1 Introduction
5.2 Liver Tumour Detection
5.2.1 Data Collection
5.2.2 Pre-Processing
5.2.3 Segmentation
5.2.4 Feature Extraction and Selection
5.2.5 Classification
5.2.6 Performance Evaluation
5.3 Discussion
5.4 Conclusion and Future Work
References
Chapter 6 Deep Learning in Photoacoustic Tomographic Image Reconstruction
6.1 Introduction
6.2 Fundamentals of Photoacoustic Tomography: Principles and Reconstruction
6.2.1 Principle of PAT
6.2.2 Optical Properties of Tissue
6.2.3 Time Domain PAT
6.2.4 Frequency Domain PAT
6.2.5 Image Reconstruction
6.3 Deep Neural Networks: Architecture, Training, Optimization
6.3.1 Architecture
6.3.2 Training
6.3.3 Optimization
6.4 Deep CNN for PAT Reconstruction: Architectures, Training, Data Augmentation, Performance Evaluation
6.4.1 Architectures
6.4.2 Training
6.4.3 Data Augmentation
6.4.4 Performance Evaluation
6.5 Generative Adversarial Networks and Recurrent Neural Networks for PAT Imaging
6.5.1 Generative Adversarial Networks
6.5.2 Recurrent Neural Networks (RNNs)
6.6 Hybrid Models in PAT Image Reconstruction—Combining Traditional and Deep Learning Methods
6.7 Conclusion
References
Chapter 7 Design and Development of Computer-Aided Diagnosis to Detect Lung Cancer Disease by Using Intelligent Deep Learning Principle
7.1 Introduction
7.2 Related Works
7.3 Methodology
7.4 Image Preprocessing
7.5 Image Augmentation
7.6 Feature Extraction
7.7 Artificial Neural Network
7.8 FPSO Technique
7.9 Hybrid ANN-PSO
7.10 Results and Discussion
7.10.1 Implementation Detail
7.10.2 Experimental Setup
7.11 Experimental Result
7.12 Conclusion and Future Scope
References
Chapter 8 Novel Methodology to Predict and Classify Liver Diseases Based on Hybrid Deep Learning Strategy
8.1 Introduction
8.2 Related Works
8.3 Methodology
8.3.1 Data Acquisition
8.3.2 Data Preprocessing
8.3.3 Feature Extraction
8.3.4 Classification Model—LSTM
8.3.5 Convolutional Neural Network
8.3.6 Hybrid Deep Learning Strategy
8.3.7 Antlion Optimization
8.4 Results and Discussion
8.4.1 Performance Evolution
8.5 Discussion
8.5.1 Comparison with Other Models
8.6 Conclusion and Future Scope
References
Chapter 9 Improvements in Analyzing Biomedical Signals and Medical Images Using Deep Learning
9.1 Introduction
9.2 Methods of Processing Biomedical Signals
9.3 Understanding Deep Learning and Its Applications in Healthcare
9.4 Significance of Biomedical Signal and Medical Image Analysis
9.5 Overview of Biomedical Signal Types
9.6 Electromyogram
9.7 Other Biomedical Signals
9.8 Challenges and Limitations in Traditional Approaches
9.9 Advancement in Deep Learning Methods
9.10 Medical Image Analysis Using Deep Learning
9.11 Conclusion
References
Chapter 10 A Survey on Lung Cancer Diagnosis Using Deep Learning Techniques
10.1 Introduction
10.1.1 Overview of Cancer in Lungs
10.1.2 Role of Deep Learning in Medical Imaging
10.1.3 Deep Learning Techniques
10.2 Deep Learning Approaches in Lung Cancer Diagnosis
10.2.1 Datasets Used
10.3 Challenges and Future Scope
10.4 Tabular Presentation of the Papers
10.5 Conclusion
References
Chapter 11 Content-Based Medical Image Retrieval Using CNN Feature Extraction and Hashing
11.1 Introduction
11.2 State of the Art
11.3 Preliminaries
11.3.1 Secure Hash Algorithm
11.4 Proposed Method
11.4.1 Preprocessing
11.4.2 Extracting the Features
11.4.3 Applying SHA-256
11.4.4 Similarity Measure
11.5 Experimental Results
11.5.1 Experiment Analysis on VIA/I-ELCAP Dataset
11.6 Conclusion and Further Studies
References
Chapter 12 Experimental Evaluation of Deep Learning-Assisted Brain Tumor Identification with Advanced Classification Methodology
12.1 Introduction
12.2 Related Works
12.3 Methodology
12.3.1 Image Preprocessing
12.3.2 Image Augmentation
12.3.3 Image Segmentation—UNET
12.3.4 GoogLeNet
12.4 Moth Flame Optimization
12.5 Results
12.6 Conclusion and Future Scope
References
Chapter 13 Study of Biomedical Segmentation Based on Recent Techniques and Deep Learning
13.1 Introduction
13.2 Overview of Techniques Used in Biomedical Segmentations
13.2.1 Segmentation Techniques
13.3 Thresholding-Based Segmentation
13.4 Region-Based Segmentation
13.5 Edge-Based/Boundary Detection
13.6 Clustering Methods
13.7 Model-Based Algorithms
13.8 Evaluation of Different Medical Image Segmentation Methods
13.9 Conclusion
References
Chapter 14 Deep CNN in Healthcare
14.1 Introduction
14.1.1 Different Types of CNN
14.2 Importance of Deep CNN
14.3 Challenges of Deep CNN
14.3.1 Volume of Data
14.3.2 Quality of Data
14.3.3 Temporality
14.3.4 Domain Complexity
14.3.5 Interpretability
14.4 Applications of Deep CNN in Healthcare
14.4.1 Medical Image Analysis
14.4.2 Disease Diagnosis and Prognosis
14.4.3 Drug Discovery and Development
14.4.4 Precision Medicine and Genomics
14.4.5 Medical Robotics and Surgical Assistance
14.5 Ethical Implications of Deep CNN
14.5.1 Safeguarding Patient Privacy and Data Security
14.5.2 Mitigating Bias and Ensuring Fairness
14.5.3 Enhancing Transparency and Interpretability
14.5.4 Clarifying Clinical Responsibility and Accountability
14.5.5 Informed Consent and Patient Autonomy
14.5.6 Preventing Overreliance on AI
14.5.7 Equitable Resource Allocation
14.5.8 Balancing Intellectual Property and Accessibility
14.5.9 Anticipating Unintended Consequences
14.5.10 Establishing Regulatory Oversight
14.6 Conclusion
References
Chapter 15 An Improved Multi-Class Breast Cancer Classification and Abnormality Detection Based on Modified Deep Learning Neural Network Principles
15.1 Introduction
15.2 Related Works
15.3 Methodology
15.3.1 Datasets
15.3.2 Data Preprocessing
15.3.3 Data Augmentation
15.3.4 EfficientNet Feature Extractor
15.3.5 Benefits of Using Efficient Net as a FeatureExtractor
15.3.6 Deep Neural Network
15.3.7 Input Layer
15.3.8 Hidden Layers
15.3.9 Neurons
15.3.10 Weights and Biases
15.3.11 Activation Functions
15.3.12 Output Layer
15.3.13 Loss Function
15.3.14 Mayfly Optimization
15.4 Results and Discussion
15.5 Conclusion and Future Scope
References
Index
