Artificial Intelligence for Smart Healthcare

Description:

This book provides information on interdependencies of medicine and telecommunications engineering and how the two must rely on each other to effectively function in this era. The book discusses new techniques for medical service improvisation such as clear-cut views on medical technologies. The authors provide chapters on communication essentiality in healthcare, processing of medical amenities using medical images, the importance of data and information technology in medicine, and machine learning and artificial intelligence in healthcare.  Authors include researchers, academics, and professionals in the field.

Preface

The healthcare sector has a huge amount of data to be processed and needs attention. Artificial intelligence and related technologies like machine learning, deep learning, IoT, big data analysis, neural networks, expert systems, physical robots, automated robotic processes, and several other technologies have the potential to revolutionize the current and future of smart healthcare by dynamically accessing information, connecting stakeholders like people, clinics/hospitals related to healthcare, and then actively managing and generating responses to the needs of the medical ecosystem in an intelligent manner. Each of these technologies has its own domain of application. The range of applications they possess is imaging analysis, patient-centric care, cloud-based network, connected staff and infrastructure, disease diagnosis and treatment, patient engagement, and others. Though the process has already begun, still a lot more needs to be done. Efficient algorithms and models that can find solutions to the above applications can break the barriers in their adoption in the healthcare sector. COVID-19 has led the medical practitioners, pharma companies, and scientists in identifying vaccines and drugs and providing optimal care to the patients. To save the world from such unprecedented times, forecasting models and prediction algorithms are the need of the hour.

This book shall to the best cover these aspects related to healthcare and also identify the challenges and find solutions for the same. Intended as a power-packed book, it shall explore not only the current state but also find effective solutions using these technologies for the future. An insightful read shall yield the audience a platform to exchange ideas and innovations by diving into the current and the future scenarios. One-to-one healing between the patient and the doctor shall change the way healthcare works.

This book is divided into three major parts. Part I: Fundamentals and Applications of AI and Enabling Technologies in Various Sectors, Part II: AI-Enabled Innovations in the Health Sector, and Part III: Security and Privacy Concerns. Part I focuses on applications of AI, deep learning, neural networks, etc. in various domains and helps the reader to gain an insight into how these technologies find their applications. Part II primarily discusses the applications of AI in the healthcare sector. Its applications for brain-related problems, disease diagnosis, smart health education, etc. are discussed in this part. Part III has contributions that discuss the security and privacy-related issues associated with it. There are 30 chapters in all. A summary of each of the parts and the chapters is provided below.

Part I: Fundamentals and Applications of AI and Enabling Technologies in Various Sectors

Chapter “A Secured Data Sharing Protocol for Minimization of Risk in Cloud Computing and Big Data in AI Application”

The study proposes a mystery allocation gathering key administration convention (SSGK) for the purpose of preventing unauthorized access to the communication measure and shared information. There’s a gathering key and a secret allocation map that distributes it. This reduces the risk of allocation information in dispersed storing and recoveries by 12%. SSGK technique with 98.25% accuracy, 97.92% sensitivity, and 93.89% recall has been established.

Chapter “Predictive Modelling for Healthcare Decision-Making Using IoT with Machine Learning Models”

An overview of machine learning techniques and their applications in the health industry is discussed. Deep learning (DL) using neural networks has lately shown promise in healthcare. A small group of people is used to anticipate how many days off employees will have. Factors such as check-in time, days away from home, and augmentation strategies in boosting the F1-score.

Chapter “Artificial Intelligence for Smart in Match Winning Prediction in Twenty20 Cricket League Using Machine Learning Model”

This study predicts the outcome of T20 matches. Cricket matches have lot of impact on the mind of the audience. A good prediction can help people accept the match’s results and avoid them putting in stress, as well as, emotional trauma. This chapter uses Machine learning techniques for a better prediction of the outcomes of the match. Several factors can be analysed including the behavioural patterns of the players. Cricketers who are self-assured typically have a high sense of mastery over their abilities, expect things to go their way and hold themselves to a higher performance standard. Mental skills such as multitasking and flexible thinking are needed to keep a cricketer’s mind from becoming overwhelmed by any complexity.

Guided athlete reflection after the performance was promoted as a valuable tool in developing and applying idiographic coping behaviors that could improve perceptions of control and self-confidence while influencing stress and emotion processes.

Chapter “Comparative Analysis of Handwritten Digit Recognition Investigation Using Deep Learning Model”

This chapter uses the CNN framework for handwritten digit recognition. It analyses and compares the performance of various conventional machine learning classifiers to the Deep learning-based Convolutional neural networks. This work can be applied in the healthcare sector, where, usually doctors are prescribing medicines through paper slips and the handwriting recognition poses a problem to the patient as well as the pharmacist. MNIST data was used to improve machine learning training and categorization.

Chapter “An Investigation of Machine Learning-Based IDS for Green Smart Transportation in MANET”

This research uses ML to distribute characteristics to IDS for MANET Green Smart Transportation. The performance of ML-IDSs may be used to detect incursions while learning about the MANET. ML optimized KDD IDS. Each controlled packet earned an anomaly score. With MANETs, datasets are scarce. The IDS for Green Smart Transportation was built using a working prototype and ML approaches. Using IDS and ML to detect and prevent MANET abuse is the approach used. This work adds to the body of information about IDSs.

Chapter “A Critical Cloud Security Risks Detection Using Artificial Neural Networks at Banking Sector”

This study focused on cloud banking developers and IT managers. The best ANN integration approach anticipates and minimizes significant security and cloud challenges. This chapter explores the use of Artificial neural networks for predicting critical cloud computing security issues in the banking sector. Optimistic predictions of significant cloud security issues can lead to better performance of the cloud-based baking sector. It holds relevance as on today in the healthcare sector, as most of the transactions made by patients are performed online. Thus, the technology proposed in the current research provides better confidence and security. Consequently, a Critical Cloud Security Risks Detection Using Artificial Neural Networks at Banking Sector would make healthcare system efficient. Optimistic projections of major cloud security issues would boost cloud-based banking performance. Performance measurements including accuracy 98.76%, sensitivity 97.34%, recall 94.53%, and F measure 97.82% were achieved. The result is better than the present methods.

Chapter “A Solution to Pose Change Challenge: Real-Time, Robust, and Adaptive Human Tracking Systems Using SURF”

The present requirement in the healthcare system is the development of smart hospitals. Reports suggest that one in twenty-five patients suffer from hospital acquire infections. It’s a major task to track down the healthcare staff, patients, visitors etc. Presently, RFID is used to track people in hospitals, and for that manual tagging of humans to be tracked is done. This is done by wearing special wristbands, gloves, or external badges to register their position at specific events of interest. This chapter presents a revolutionary GRABCUT implementation on the interest point-based method. This approach removes background descriptors from the object model, which is then utilized by a SURF-based tracker. An auto-tuned classifier is then used to distinguish between posture change and occlusion. This chapter’s algorithm is adaptable, resilient, and real time. The tests are done both inside and outdoors.

Chapter “Analysis on Identification and Detection of Forgery in Handwritten Signature Using CNN”

The Forgery of Handwritten Signature is a critical task but hackers sometimes copied the information easily. Many standard approaches like evolutionary algorithm, Random-forest optimization, Xboosting machine learning, and SVM cannot solve signature forging. The Forgery of Handwritten Signature is a critical task but hackers sometimes copied the information easily. This chapter concentrate on Identification and Detection of Forgery in Handwritten using CNN deep learning technology. The available medical with digital information has been used train the CNN architecture. The testing process has been initiated through testing block in CNN deep intelligence. The privacy, ethical and legal issues of Forgery has been identified and suggested which data is original. The document writer checks the falsified signature. Existing approaches struggle to identify fake documents and signatures. The average accuracy is 83.3%, sensitivity 99.2%, recall 98.24%, F score 97.23%, and throughput 98.91%. These enhanced findings outperformed prior models.

Chapter “Experimental Analysis of Internet of Technology-Enabled Smart Irrigation System”

The new Cloud-enabled Smart Agri-Handling Strategy (CSAHS) is presented. The CSAHS technology will allow farmers to remotely monitor their crops and get weather alerts. For this study, the IoT connects farm smart land equipment to a distant cloud server. The ARM IoT-Web Module has Wi-Fi. It uses data (like PH, soil moisture, rain, temperature, and humidity) from smart sensors to operate agricultural smart equipment. This IoT-enabled agricultural gadget saves water. The outcome is optimum crop irrigation at the right time. The CSAHS reduces labor, reduces agricultural losses, and provides timely weather information.

Chapter “Analysis on Exposition of Speech Type Video Using SSD and CNN Techniques for Face Detection”

The use of multimedia has led to more and more videos being generated. Thus, their storage, management, and mapping becomes a serious problem. This paper employs a two step process for emotions detection. Face Detection is done using SSD (Single Shot Multibox Detector) and emotion classification is done using CNN. Face-SSD employs fully CNN to identify video uploads. This study used CNN-based speech type video extraction and face identification to estimate storage and simplify content indexing. Finally, performance metrics like 98.45% accuracy, 97.34% sensitivity, 94.23% recall, and throughput have been calculated

Part II: AI-Enabled Innovations in the Health Sector

Chapter “Depressive Disorder Prediction Using Machine Learning-Based Electroencephalographic Signal”

Early detection of depression is crucial. This chapter uses EEG signals from a publicly available database to evaluate sleep disorders and alcoholism. Its classifier tools may assist classify topics with disorders. Frequency bands (Apha, Delta, Theta, and Beta) provide the parameters.

Chapter “Generation of Masks Using nnU-Net Framework for Brain Tumor Classification”

By dealing with the dataset assortment seen in the medical domain datasets, this paper proposes a technique called nnU-Net with adjustments in the encoder-decoder design. The MRI dataset is used to mine the shape from the radiometric properties. Pre-processing, network architecture, training, and post-processing for every new job and pragmatic choices are automatically configured.

Chapter “A Brain Seizure Diagnosing Remotely Based on EEG Signal Compression and Encryption: A Step for Telehealth”

The need for e-health and remote health is increasing andrequires information to be transmitted between the patients and the healthcare stakeholders. as such,data compression and encryption becomes crucial. This work presents both EEG data reduction and encryption techniques. The Huffman adaptive coding and encryption technique using 256-bit AES (CBC) is performed. It improves compression and encryption performance on the CHB-MIT Scalp EEG database.

Chapter “A Deep Convolal Neural Network-Based Heart Diagnosis for Smart Healthcare Applications”

Coronary artery diseases were classified using hybrid filtering (HCAD). A probabilistic Adaptive Random Forest classifier was utilized to predict HCAD data. This study classified CTA/heart images using level set formulation and GHSB (Gaussian Hue, Saturation, and Brightness). For cardiac disorders, GHSB employs fusion. This method is slow and inaccurate. Adaptive filters and threshold segmentation can identify HCAD faster. It increases True positives using adaptive machine learning. Heart disease is predicted by RFO segmentation and classification. Application of DCNN-based feature selection and image net classification in healthcare is necessary for healthcare applications. This research employs GHSB and RFO models on MRI cardiac scans. This model finds and extracts the CAB from the image. Thus, the blockade’s breadth and length are determined. Last but not least, 99.74% accuracy, 98.81% precision, 98.14% recall, 94.87% F1 score, 59.26% PSNR, 0.0989% CC, and 99.45% sensitivity were attained.

Chapter “A Dynamic Perceptual Detectors Module-Related Telemonitoring for the Intertubes of Health Services”

An evaluation of intellectual data approaches for computer systems with restricted resources is investigated by dynamically configuring the vehicle’s hardware and software to the required operating mode. To test this method, a low-power microcontroller and a neural network model are used. Using the MIT-BIH Arrhythmia testbed, adapting the node design to the strain during execution reduces power use by 50%. A quantized neural network could detect arrhythmias with 98% accuracy.

Chapter “ConvNet-Based Deep Brain Stimulation for Attack Patterns”

An algorithm for identifying Parkinson’s disease using limited cross-sectional brain structural MRI data was developed and validated in this work. To encode prediction and classification, a type of RNN called LSTM may learn order dependence. Computational linguistics and voice recognition are two examples. Deep learning LSTMs are challenging. An LSTM predicts the highest trembling speed of fake and real stimuli. Assault patterns were recognized by the architecture, alerting the doctor.

Chapter “Web-Based Augmented Reality of Smart Healthcare Education for Machine Learning-Based Object Detection in the Night Sky”

The authors suggested Space360, a virtual planetarium for people who cannot go to the planetarium due to COVID. Space360 allows users to observe every celestial body in the night sky at any time. Space360 is about stargazing and seasonal constellations and provides knowledge on the celestial body. This feature will display details such as visibility, declination, etc. This applies to all 88 visible constellations and eight planets in our solar system.

Chapter “The Role of Augmented Reality and Virtual Reality in Smart Health Education: State of the Art and Perspectives”

AR and VR can improve teaching by providing immersive multimodal environments with different sensory characteristics. The chapter finishes with several novel approaches to challenges and research topics for future educators interested in these new technologies.

Chapter “Estimation of Thyroid by Means of Machine Learning and Feature Selection Methods”

This study uses KNN, naive Bayes, and SVM to categorize thyroid datasets. Comparing several machine learning algorithms helps forecast illness. The E.S.T. D.D. Model is better for future thyroid ailment detection. Existing applications were not suitable for pre-stage thyroid diagnosis. So, a complex model is necessary. Accuracy of 98.53%, recall of 97.23%, the throughput of 98.34%, and sensitivity of 99.23% were attained.

Chapter “A Multiuser-Based Data Replication and Partitioning Strategy for Medical Applications”

This research provides a real-time cloud computing hybrid multi-user data replication and partitioning solution for medical applications. Two stages are used in this study: data replication and multi-user partitioning. To retrieve data, numerous servers’ machine decision patterns are replicated. This stage partitions cloud data. Data replication and partitioning outperform previous solutions on large medical decision patterns. A wide range of replicant protocols exist. A review of distributed storage and content management system replication approaches is presented. The test has 98.23% accuracy, 94.53% sensitivity, and 92.29% recall.

Chapter “A Deep Study on Thermography Methods and Applications in Assessment of Various Disorders”

This research examines numerous thermography methods and their usefulness. The temperature of an object influences its infrared emission. Thermal imaging is utilized in medical diagnostics, mechanical and electrical maintenance, as well as building heat loss measurement. This chapter concentrates on temperature measurement and non-destructive testing. An overview of infrared thermography, temperature measurement, and non-destructive testing is provided. Recent advances in these fields are also addressed. This study’s accuracy was 98.67%, sensitivity was 94.45%, and recall was 92.34%.

Chapter “A Smart Healthcare Cognitive Radio System for Future Wireless Commutation Applications with Test Methodology”

Determining the whole system rather than simply individual components is challenging due to the lack of a device-independent design. To overcome this problem, this study proposes the CORATM Cognitive Radio Research Technique using MICR. MICR advises PU and SU to use behavior-based node assessment to assess cognition. Assessing both main and secondary user success may be highly beneficial. This study’s results include accuracy 98.34%, sensitivity 98.56%, recall 98.51%, F scores 98.65%, and throughput 99.32%. It competes with 4G and 5G.

Chapter “Indication of COVID-19 and Inference Employing RFO Classifier”

Pandemic origins have left many jobless in developing countries and beyond. Managing the GDP is challenging. Obvious limitations must be addressed in order to determine COVID-19 impact. So, we employ the RFO machine learning model. The model categorizes variables affecting unemployment. This performance measure attained metrics like accuracy 98.45%, recall 97.235%, sensitivity 95.345%, Fmeasure 93.785%, and sensitivity 99.325%. These measures outperform techniques in data sensitivity estimations. The RFO-based COVID-19 unemployment estimates social and ethical characteristics. This application tests machine learning techniques in today’s economy.

Chapter “Magnetic Resonance Images for Spinal Cord Location Detection Using Deep Learning Model”

Deep learning is used to predict MRI images of spinal cord damage. The suggested deep learning model is CONN; the layers are max-pooling, hidden, dense,

normalized, and ReLu. This model has 13 layers and performed better than previous models. 97.23% recall, 95.12% precision, and 98.71% accuracy were achieved.

Chapter “COVID-19 Recognition in X-Ray and CTA Images Using Collaborative Learning”

COVID-19 is a serious illness that may be fatal. This research proposes Otsu thresholding under random forest optimization machine learning to diagnose COVID-19 X-ray images. Finally, estimation accuracy 99.46%, sensitivity 99.48%, recall 99.36%, and F1 score of 99.42% were achieved.

Chapter “Artificial Intelligence for Enhancement of Brain Image Using Semantic Segmentation CNN with IoT Classification Techniques”

The major goal is to use a medical database to show how various brain imaging variables affect malignancies. Methods enhance data segmentation, feature extraction, and classification. To assess the proposed semantic segmentation, deep convolution neural network classification algorithms are used. This approach and medical data model may be scalable according to experimental results.

Part III: Security and Privacy Concerns

Chapter “A Novel Framework for Privacy Enabled Healthcare Recommender Systems”

The authors suggested a two-tier structure to explore it in this study. The patient’s identity is first anonymized, then data is randomly generated using homomorphic encryption. The presented framework serves two purposes. It is free of third-party participation and promotes privacy by integrating two-layer design for patient data protection. Also, the proposed solution may aid as a standard for protecting patients’ privacy and integrity.

Chapter “An Evaluation of RSA and a Modified SHA-3 for a New Design of Blockchain Technology”

This research added a new criterion for evaluating the RSA algorithm used in blockchain technology. To improve hashing speed, a redesigned hash algorithm (SHA3) was developed by replacing AND and NOT by ADD. The major result was

less memory use and faster RSA method complexity and hashing. Finally, compression between proposed algorithms (amendment RSA and modified SHA-3) improved memory, security, and speed.

Chapter “Quality of Smart Health Service for Enhancing the Performance of Machine Learning-Based Secured Routing on MANET”

The goal is to enhance routing performance for ad hoc networks utilizing multipath and a hop count-based routing parameter. Better routing services for congested applications, particularly multimedia apps can be achieved through this model. On the CSMA/CA approach, EED and PLR are evaluated in MAC layer associations. These variables are used with application QoS needs to build additional routing parameters. Typical multipath protocols avoid application-layer route preferences under varying congestion circumstances by using constant routing parameters. The suggested approach outperforms the AOMDV standard in terms of PDR, congested overheads, and EED for application congestion types.

Chapter “Spoof Attacks Detection Based on Authentication of Multimodal Biometrics Face-ECG Signals”

Face and ECG biometric authentication is the purpose of this project. Face and ECG multimodality system has the best statistical performance. ECG has been shown to improve the anti-spoofing capabilities of conservative biometric-based systems. It describes an ECG-based authentication mechanism. Awica Wavelet Transformation techniques are required alongwith an improved method of authentication. The results were 94% accuracy, 98% FAR, 7.93% FRR, and 14.99-minute process time.

This book intends to provide the reader with extensive coverage of AI, its enabling technologies in the healthcare sector. We wish to thank whole-heartedly the entire team of Springer, particularly Eliška Vlčková, Managing editor, EAI, for her constant support and guidance. We also extend our thanks to all the contributors and reviewers. They are the major stakeholders in the editing of a book. Our gratitude to our family and friends. Last but not the least, we bow before the Almighty for his blessings throughout the journey.

Preface
Part I: Fundamentals and Applications of AI and Enabling Technologies in Various Sectors
Chapter “A Secured Data Sharing Protocol for Minimization of Risk in Cloud Computing and Big Data in AI Application´´
Chapter “Predictive Modelling for Healthcare Decision-Making Using IoT with Machine Learning Models´´
Chapter “Artificial Intelligence for Smart in Match Winning Prediction in Twenty20 Cricket League Using Machine Learning Mode…
Chapter “Comparative Analysis of Handwritten Digit Recognition Investigation Using Deep Learning Model´´
Chapter “An Investigation of Machine Learning-Based IDS for Green Smart Transportation in MANET´´
Chapter “A Critical Cloud Security Risks Detection Using Artificial Neural Networks at Banking Sector´´
Chapter “A Solution to Pose Change Challenge: Real-Time, Robust, and Adaptive Human Tracking Systems Using SURF´´
Chapter “Analysis on Identification and Detection of Forgery in Handwritten Signature Using CNN´´
Chapter “Experimental Analysis of Internet of Technology-Enabled Smart Irrigation System´´
Chapter “Analysis on Exposition of Speech Type Video Using SSD and CNN Techniques for Face Detection´´
Part II: AI-Enabled Innovations in the Health Sector
Chapter “Depressive Disorder Prediction Using Machine Learning-Based Electroencephalographic Signal´´
Chapter “Generation of Masks Using nnU-Net Framework for Brain Tumor Classification´´
Chapter “A Brain Seizure Diagnosing Remotely Based on EEG Signal Compression and Encryption: A Step for Telehealth´´
Chapter “A Deep Convolal Neural Network-Based Heart Diagnosis for Smart Healthcare Applications´´
Chapter “A Dynamic Perceptual Detectors Module-Related Telemonitoring for the Intertubes of Health Services´´
Chapter “ConvNet-Based Deep Brain Stimulation for Attack Patterns´´
Chapter “Web-Based Augmented Reality of Smart Healthcare Education for Machine Learning-Based Object Detection in the Night S…
Chapter “The Role of Augmented Reality and Virtual Reality in Smart Health Education: State of the Art and Perspectives´´
Chapter “Estimation of Thyroid by Means of Machine Learning and Feature Selection Methods´´
Chapter “A Multiuser-Based Data Replication and Partitioning Strategy for Medical Applications´´
Chapter “A Deep Study on Thermography Methods and Applications in Assessment of Various Disorders´´
Chapter “A Smart Healthcare Cognitive Radio System for Future Wireless Commutation Applications with Test Methodology´´
Chapter “Indication of COVID-19 and Inference Employing RFO Classifier´´
Chapter “Magnetic Resonance Images for Spinal Cord Location Detection Using Deep Learning Model´´
Chapter “COVID-19 Recognition in X-Ray and CTA Images Using Collaborative Learning´´
Chapter “Artificial Intelligence for Enhancement of Brain Image Using Semantic Segmentation CNN with IoT Classification Techn…
Part III: Security and Privacy Concerns
Chapter “A Novel Framework for Privacy Enabled Healthcare Recommender Systems´´
Chapter “An Evaluation of RSA and a Modified SHA-3 for a New Design of Blockchain Technology´´
Chapter “Quality of Smart Health Service for Enhancing the Performance of Machine Learning-Based Secured Routing on MANET´´
Chapter “Spoof Attacks Detection Based on Authentication of Multimodal Biometrics Face-ECG Signals´´
Contents
Part I: Fundamentals and Applications of AI and Enabling Technologies in Various Sectors
A Secured Data Sharing Protocol for Minimisation of Risk in Cloud Computing and Big Data in AI Application
1 Introduction
1.1 Cloud Storage for Big Data
1.2 Techniques
1.3 Extraction
2 Existing Methods
3 Proposed System
3.1 Module Implementation
3.1.1 Cloud Provider
3.1.2 Data Owner
3.1.3 The Assembly Members
4 Results and Analysis
5 Conclusion
References
Predictive Modelling for Healthcare Decision-Making Using IoT with Machine Learning Models
1 Introduction
2 Related Works
3 Machine Learning Models and Classification in Healthcare Applications
3.1 Supervised Machine Learning
3.2 Unsupervised Learning
3.3 Semi-supervised Learning
3.4 Reinforcement Learning
4 Healthcare Applications of ML in Diagnosis
5 Secure and Privacy-Preserving Use of Healthcare Measures on IoT and ML
5.1 Sources of Vulnerabilities in ML
5.2 Vulnerabilities Due to Data Annotation
5.3 Vulnerabilities in Model Training
6 Results and Discussion
6.1 Healthcare Diagnoses Outcome
6.2 Feature Predictions
7 Conclusions and Future Work
References
Artificial Intelligence for Smart in Match Winning Prediction in Twenty20 Cricket League Using Machine Learning Model
1 Introduction
2 Related Works
3 Proposed Methodology
3.1 Machine Learning Classification
3.2 Supervised Learning (SL)
3.3 Unsupervised Learning (UL)
3.4 Reinforcement Learning (RL)
4 Result and Discussions
4.1 Linear Regression (LR)
4.2 Step-by-Step Process of Machine Learning Classifications
5 Conclusion
References
Comparative Analysis of Handwritten Digit Recognition Investigation Using Deep Learning Model
1 Introduction
1.1 Abbreviations and Acronyms
1.1.1 Units
2 Related Works
3 Proposed CNN Image Classification
3.1 PReLU
3.2 Shrinking
3.3 Non-linear Mapping
3.4 Expanding
3.5 Deconvolution
4 Mathematical Model
4.1 Subsampling Layer
5 Result and Discussion
6 Conclusion
References
An Investigation of Machine Learning-Based IDS for Green Smart Transportation in MANET
1 Introduction
2 Related Works
3 Proposed ML-Based Green Smart Transportation IDS
3.1 Pre-processing and Data Augmentation
3.2 Parameters Optimization
3.3 Ensemble Learning
3.4 Algorithm 1 for ML Classification for Malicious Node Detection
3.5 Algorithm 2 for Optimal Solution for Acceptable Error Rate
3.6 Algorithm 3 for Packet Anomaly Prediction
4 Mathematical Model of ML Process
5 Result and Discussion
5.1 Simulations
5.1.1 Findings of KDD
6 Performance Analysis
7 Conclusion
References
A Critical Cloud Security Risks Detection Using Artificial Neural Networks at Banking Sector
1 Introduction
2 Artificial Neural Networks
3 Literature Survey
3.1 Cloud Computing Modeling Concepts for Banking Organizations
3.2 Cloud Security Issues
4 Methodology (Materials and Methods)
4.1 Neural Network
5 Outcomes and Discussion
6 Conclusion
References
A Solution to Pose Change Challenge: Real-Time, Robust, and Adaptive Human Tracking Systems Using SURF
1 Vision-Based Tracking
1.1 Difficulties in Visual Tracking
1.2 Required Features of Visual Tracking
1.3 Feature Descriptors for Visual Tracking
1.4 Online Learning Algorithms
1.5 Applications
2 Object Detection and Tracking
2.1 Object Representation
2.2 Object Detection
2.3 Object Tracking
2.4 Prediction Methods
3 Literature Review
4 Motivation
5 Terminology
5.1 Categorization
6 Evaluation Metrics
7 SURF-Based Algorithm to Deal with Pose Change Challenge
8 Problem Statement
9 Tracking Algorithm
10 Results
11 Conclusion
References
Analysis on Identification and Detection of Forgery in Handwritten Signature Using CNN
1 Introduction
2 Literature Review
3 Methodology
3.1 Convolution Neural Network (CNN)
3.1.1 Convolution Filtering
3.1.2 Implementation
3.1.3 Flow Diagram
3.1.4 Data Acquisition
3.1.5 Preprocessing
3.1.6 Gray to Binary
3.1.7 Noise Removal and Resizing
3.1.8 Adding CNN Layers
3.1.9 Pooling Layer
3.1.10 Flatten
3.1.11 Dense- Softmax
3.1.12 Feature Extraction
4 Results
5 Discussions
6 Conclusion
6.1 Future Scope
References
Experimental Analysis of Internet of Technology-Enabled Smart Irrigation System
1 Introduction
2 Related Study
3 Methodologies
3.1 pH Sensor
3.2 Soil Moisture Sensor
3.3 Rain Sensor
3.4 Temperature and Humidity Sensor
3.5 Cloud-Enabled Smart Agri-Handling Strategy (CSAHS)
3.6 Cloud Optimization Strategy
4 Discussion
5 Conclusion and Future Scope
References
Analysis on Exposition of Speech Type Video Using SSD and CNN Techniques for Face Detection
1 Introduction
2 Related Work
3 Approach
3.1 Single Shot Multi-box Detector (SSD)
3.2 Convolution Neural Network (CNN)
3.3 Frame Selection
3.4 VSUMM
3.5 Video Exposition
4 Results and Discussion
5 Conclusion
References
Part II: AI-Enabled Innovations in the Health Sector
Depressive Disorder Prediction Using Machine Learning-Based Electroencephalographic Signal
1 Introduction
2 Related Works
3 Scope of the Work
3.1 The Research Objective
4 Proposed Block Diagram
4.1 Overall Working Principle of EEG Signal
4.2 Procedure for PSD Calculation
4.3 Algorithm for Feature Extraction by ML Method
5 Results and Discussion
6 Conclusion and Future Work
References
Generation of Masks Using nnU-Net Framework for Brain Tumour Classification
1 Introduction
1.1 Imaging Modalities
1.2 General Analysis Objectives
2 Literature Review
3 Proposed Methodology
3.1 Materials and Methods
3.2 Experimental Setup
3.3 Model Training and Making Masks
4 Conclusion
References
A Brain Seizure Diagnosing Remotely Based on EEG Signal Compression and Encryption: A Step for Telehealth
1 Introduction
2 Related Literature Works
3 Methodology
4 Research Parameters
4.1 Exchange of Diffie-Hellman Key
4.2 Signal Block Creation
4.3 Adaptive Huffman Encoding (AdHuEn)
5 Hybrid Cryptography for EEG Signal
6 Findings and Discussion
6.1 Compression Performance
6.2 Analysing the Security of EEG Signals
7 Conclusion
References
A Deep Convolutional Neural Network-Based Heart Diagnosis for Smart Healthcare Applications
1 Introduction
2 Literature Survey
3 Methodology
4 Results and Discussion
5 Conclusion
References
A Dynamic Perceptual Detector Module-Related Telemonitoring for the Intertubes of Health Services
1 Introduction
2 Related Work
3 Data Transmission Layout Adaptation
4 Node Structure for IoMT
4.1 Platform Surface of Hardware
4.2 OS/Layer of the Load Balancer
4.2.1 FreeRTOS
4.2.2 Cortex Microcontroller
5 Approach Model
6 Assistance for Flexibility: The ADAptive Environment Scheduler
7 Use-Case Evaluation
7.1 Information Collected Is the First Intraoperative Phase
7.2 Saturation Identification Is the Client Operating Mode
7.3 Analysis of CNN o.m. in the Mode of Operation 3
8 Peak Detection
9 Activation for Neural Networks
9.1 Quantization
9.2 Investigation of Archetypes
9.3 Amplification
10 Research Outcomes
10.1 Assessment of Energy Utilization
10.2 Indicators of Electricity Consumption
10.2.1 Instance: 50 bpm
10.2.2 Instance: 100 bpm
10.2.3 Instance: 200 bpm
10.3 Estimation of Energy Usage Related to Power Model and Operation Mode
11 Conclusion
References
ConvNet-Based Deep Brain Stimulation for Attack Patterns
1 Introduction
1.1 Planning and Positioning of DBS
1.1.1 The Objective of Machine Learning (ML) Model
2 Literature Survey
2.1 Findings of Machine Learning in DBS
3 Proposed Work
3.1 Objectives of Machine Learning
3.2 Input Signal Collection
3.3 Class Separation
3.4 Rest Tremor Velocity
3.5 Parameter Setting
3.6 Convolutional Neural Network
3.7 Performance Measures for Precision, Error Rate, and Fully Connected Layers
4 Experimental Results
4.1 Software Environment
5 Conclusion
5.1 Future Work
References
Web-Based Augmented Reality of Smart Healthcare Education for Machine Learning-Based Object Detection in the Night Sky
1 Introduction
2 Related Works
3 Materials and Methods
3.1 Sky Map
3.2 Space Catalogue
3.3 Experimental Setup
3.4 Validation
3.5 ML and OpenCV to Detect Objects
4 Dataset
4.1 Data Preparation
4.2 Metrics
5 Conclusion
References
The Role of Augmented Reality and Virtual Reality in Smart Health Education: State of the Art and Perspectives
1 Introduction
2 Augmented Reality (AR) in Education
3 Virtual Reality (VR) in Education
4 Recent Advancements of AR/VR in Education
5 Challenges of Adopting AR/VR in Education
6 Future Scope of AR/VR in Education
7 Conclusion
References
Estimation of Thyroid by Means of Machine Learning and Feature Selection Methods
1 Introduction
1.1 A General Overview of Thyroid Disease
1.1.1 The Role of the Thyroid in the Body´s Health
1.1.2 Hyperthyroidism
1.1.3 Thyroid Hormones
2 Literature Survey
2.1 Problem Declaration
2.1.1 Existing System
Limitations of Current Material
2.1.2 Proposed System
2.2 System Overview
2.3 Machine Learning Implementation
2.3.1 SVM Algorithm (Support Vector Machine)
2.3.2 KNN Algorithm (K-Nearest Neighbor)
2.3.3 NB (Naiïve Bayes)
2.3.4 The Dataset Description
2.4 Modules
2.4.1 Admin
2.4.2 User
3 Result and Analysis
4 Conclusion
References
A Multiuser-Based Data Replication and Partitioning Strategy for Medical Applications
1 Introduction
2 Related Works
2.1 Partitioning Strategies
2.2 Proposed Model
3 Conclusion
References
A Deep Study on Thermography Methods and Applications in Assessment of Various Disorders
1 Introduction
2 Materials and Methods
3 Literature Survey
3.1 Proposed System
4 Result Analysis
5 Conclusion
References
A Smart Healthcare Cognitive Radio System for Future Wireless Commutation Applications with Test Methodology
1 Introduction
1.1 MICR
2 Related Work
2.1 Cognitive Radio Definitions and Architectures
2.2 Numerical Analysis
2.3 Cognitive Radio Test Methodologies
3 Proposed Test Methodology
3.1 Overview
3.2 Benchmarks
3.3 Test Outline
4 Methodology
4.1 Approach
4.2 Evaluation Technique
4.3 Performance Metrics
4.4 Workload
4.5 System Under Test
4.6 Experimental Design
5 Results and Analysis
5.1 Effect of SU CE on SU Throughput and BER
5.2 Throughput vs BER in MICR
5.3 Scoring
6 Conclusion
References
Indication of COVID-19 and Inference Employing RFO Classifier
1 Introduction
1.1 COVID-19 and Its Effecting Elements
1.2 Immunity Issues
1.3 Vaccine and Its Working
2 Symptoms and Samples
2.1 Prevention
2.2 Diagnosis Process
2.3 RFO Methodology
3 Results and Discussion
4 Conclusion
References
Magnetic Resonance Images for Spinal Cord Location Detection Using a Deep-Learning Model
1 Introduction
2 Literature Review
3 Related Works
4 Proposed Work
4.1 Spinal Cord Centerline Recognition
4.2 Spinal Cord and Mean Segmentation
4.3 Implementation
4.4 Evaluation
4.5 Spinal Cord Centerline Recognition
4.6 Spinal Cord Segmentation
4.7 Mean Segmentation of Lesions
4.8 Inter-Rater Variability of the Mean Lesion Segmentation
5 Results
6 Conclusion
References
COVID-19 Recognition in X-RAY and CTA Images Using Collaborative Learning
1 Introduction
2 Limitations of Earlier Models
3 Methodology
3.1 System Architecture
3.2 Dataset Details
3.3 Network Training Configuration
3.4 CNN Founded Feature Separator and Classification
3.5 Segmentation of the COVID-19-Affected Region
4 Results
5 Conclusion
References
Artificial Intelligence for Enhancement of Brain Image Using Semantic Segmentation CNN with IoT Classification Techniques
1 Introduction
2 Literature Survey
3 Proposed System
3.1 Semantic Segmentation
4 Result and Discussion
5 Conclusion
References
Part III: Security and Privacy Concerns
A Novel Framework for Privacy Enabled Healthcare Recommender Systems
1 Introduction
2 Materials and Methods
2.1 Background
2.1.1 Users Health Data Concerns
2.1.2 Data Breaches: US Data (2010-2020)
2.2 Privacy Aspects in Healthcare System
2.3 Existing Privacy Enabled Healthcare Recommender System
3 Results
3.1 Proposed Healthcare Recommender System
3.1.1 Steps of Recommendation Generation Through Our Framework
4 Discussion
5 Conclusion
References
An Evaluation of RSA and a Modified SHA-3 for a New Design of Blockchain Technology
1 Introduction
2 Related Work
3 SHA-3 and RSA Algorithm
3.1 SHA-3
3.2 RSA Algorithm
4 Proposed System
4.1 First Part: The Client
4.2 Second Part: The Server
4.3 Improve Modification RSA Algorithm (IM_RSA)
4.4 Lightweight Hash Function
5 Result and Discussion
6 NIST Statistical Suite Tests
7 Conclusion
References
Quality of Smart Health Service for Enhancing the Performance of Machine Learning-Based Secured Routing on MANET
1 Introduction
2 Related Works
2.1 Cross Layered QoS Routing
2.2 Assessing the Rate of Packet Losses
3 Assessing Connection Delays
4 Purpose of Routing Parameters
5 QoS Forwarding
6 Performance Analysis
6.1 Results and Discussion
6.2 Average Delays
6.3 Network Throughput
7 Rate of Packet Delivery
8 Congestion Overheads
9 Conclusion
References
Spoof Attacks Detection Based on Authentication of Multimodal Biometrics Face-ECG Signals
1 Introduction
2 Review of Related Literature
3 Biometric System Attacks Patterns
3.1 Direct Attack Pattern
3.2 Indirect Attack Pattern
4 Current Techniques and Its Restrictions
5 The Proposed Techniques and Its Benefits
5.1 Face Based Biometric Authentication
5.2 Facial Recognition Data Collection (FRDC)
5.3 ECG Signals-Based Biometric Authentication
5.4 ECG Signal Acquisition
6 Resist Attacks Techniques (RAT)
6.1 Vitality Detecting Method (VDM)
6.2 Cryptosystems Biometric
6.3 Avoid the Artefacts
6.4 Artefacts Cancelation
6.5 Channel Selection
7 Feature Extracting
8 Frequency Domain Approaches
9 Data Reduction
10 Classification
11 Face and ECG Fusion
11.1 Proposed MMB Authentication Using Face and ECG
12 Results and Outcomes
13 Conclusion
References
Index