Machine Vision and Augmented Intelligence: Select Proceedings of MAI 2022

Description

This book comprises the proceedings of the International Conference on Machine Vision and Augmented Intelligence (MAI 2022). The conference proceedings encapsulate the best deliberations held during the conference. The diversity of participants in the event from academia, industry, and research reflects in the articles appearing in the book. The book encompasses all industrial and non-industrial applications. This book covers a wide range of topics such as modeling of disease transformation, epidemic forecast, image processing, and computer vision, augmented intelligence, soft computing, deep learning, image reconstruction, artificial intelligence in health care, brain-computer interface, cybersecurity, social network analysis, and natural language processing.​ Contents About the Editors Detection of Physical Impairments on Solar Panel Using YOLOv5 1 Introduction 2 Image Pre-Processing Parameters and Algorithm 2.1 Collection of Specimen Images Data 2.2 Detection of Object Using YOLOv5 Algorithm 2.3 Data Augmentation for Accuracy Improvement 3 Training and Loss Function of Datasets 4 Evolution Metrics for Damage Detection 4.1 Object Detection Using Intersection Over Union (IoU) 4.2 Precision of Prediction of Damage Class 4.3 Recall of Total Relevant Result 4.4 Mean Average Precision (mAP) of Precision–Recall Curve 5 Training Results 6 Conclusion and Future Work References Image Processing Techniques on Porous Silicon to Estimate Porosity and Pore Size 1 Introduction 2 Image Processing of Porous Silicon SEM Images 2.1 Image De-noising and Filtration to Reduce Error 2.2 Morphological Operations of Images 2.3 Image Thresholding 3 Parameter Selection for Noise Filtration and Closing Operation 4 Results and Discussions 5 Conclusion References Solar PV System Fault Classification Using Machine Learning Techniques 1 Introduction 2 Simulation of the 1.3 KW PV-System for Synthetic Data Generation 2.1 Healthy Condition 2.2 Shading Faults 2.3 Line-To-Line Fault 3 Machine Learning Techniques 3.1 Accuracy Check and Results 3.2 Prediction Results 4 Conclusion References A Lightweight Network for Detecting Pedestrians in Hazy Weather 1 Introduction 2 Related Work 3 Proposed Method 4 Experiments and Results 5 Conclusion References Hand Gesture-Based Recognition System for Human–Computer Interaction 1 Introduction 1.1 Novelty of Work 2 Literature Review 3 Methodology 3.1 Basic Overview 3.2 Step-By-Step Process 3.3 Tools Used 4 Result Analysis 5 Conclusion and Future Scope 5.1 Summary of Work 5.2 Conclusion 5.3 Future Scope References An Overview of Machine Learning Techniques Focusing on the Diagnosis of Endometriosis 1 Introduction 2 Image Datasets Used 3 Performance Metrics 3.1 Accuracy 3.2 Sensitivity 3.3 Specificity 3.4 Precision 3.5 F1-Score 3.6 AUC 3.7 AUROC 3.8 P-Value 3.9 C-Index 4 Techniques 4.1 Unsupervised Machine Learning [18] 4.2 Logistic Regression [19] 4.3 Logistic Regression + Naive Bayes [20] 4.4 Classification and Regression Trees [21] 4.5 Computer Vision [22] 4.6 EXtreme Gradient Boosting (XGB) [24] 4.7 Natural Language Processing [25] 4.8 Decision Tree [13] 4.9 Decision Tree + Generalized Linear Model [26] 4.10 Convolutional Neural Network [27] 4.11 ResNet50 Convolutional Neural Network [7] 4.12 VGGNet-16 Model [16] 4.13 Artificial Neural Network [21] 4.14 Deep Neural Network [14] 4.15 Deep Learning Along with Histopathological Subtypes [33] 5 Comparison of Several Machine Learning Techniques 6 Conclusion References A Time-Dependent Mathematical Model for COVID-19 Transmission Dynamics and Analysis of Critical and Hospitalized Cases with Bed Requirements 1 Introduction 2 Proposed Methodology 2.1 A Time-Dependent SEAIHCRD Model 3 Data Analysis and Parameter Estimation 4 Results and Discussion 4.1 A Time-Dependent SEAIHCRD Model for Brazil 4.2 A Time-Dependent SEAIHCRD Model for India 4.3 A Time-Dependent SEAIHCRD Model for Mexico 4.4 A Time-Dependent SEAIHCRD Model for Russia 4.5 A Time-Dependent SEAIHCRD Model for South Africa 4.6 A Time-Dependent SEAIHCRD Model for the United States 5 Conclusion References A Compartmental Mathematical Model of COVID-19 Intervention Scenarios for Mumbai 1 Introduction 2 Proposed Methodology 3 Data Analysis and Parameter Estimation 4 Results and Discussion 4.1 Proposed Model for Mumbai 4.2 Performance Evaluation Criteria 5 Summary References A Mathematical Model for the Effect of Vaccination on COVID-19 Epidemic Spread 1 Introduction 2 Proposed Method 3 Mathematical Model 4 Result and Discussion 5 Summary References Text Classification Using Hybridization of Meta-Heuristic Algorithm with Neural Network 1 Introduction 2 Literature Review 3 Proposed Model 3.1 Pre-processing 4 Feature Selection 4.1 Sentiment Classification 5 Experiments and Result 5.1 Dataset Utilized for Experiment 5.2 Comparative Analysis of the Proposed Algorithm and Existing Research 6 Conclusions References GAN-Based Data Generation Technique and its Evaluation for Intrusion Detection Systems 1 Introduction 2 Related Work 2.1 Feature Selection Method 2.2 Machine Learning and Deep Learning Methods for Intrusion Detection System: 2.3 Classifiers Used for Anomaly Detection 3 System Architecture 4 Proposed Work 4.1 Dataset Description 4.2 Dataset Collection 4.3 Pre-processing Phase 4.4 Model Description 5 Result Analysis 5.1 Performance Comparison with Other Dimensionality Reduction Techniques 5.2 Performance of Different ML Techniques on Generated Data Samples 6 Conclusion References VSCM: Blockchain-Based COVID-19 Vaccine Supply Chain Management 1 Introduction 2 Contribution 3 Literature Review 4 Preliminaries 4.1 Blockchain 4.2 Smart Contract 4.3 Cold Chain Network 5 Proposed Model 5.1 Proposed Blockchain-Based Vaccine Distribution System 5.2 Description of Three Blockchain in Our Proposed Model 5.3 Step Wise Workflow of Our Proposed Model 5.4 Security Analysis 6 Performance Evaluation 7 Conclusion References Polyp Segmentation Using Efficient Multi-supervision Net: A Deep Learning Technique Uses Attention Unit and EfficientNet Model 1 Introduction 2 Related Work 3 Experimental Results and Discussion 3.1 Dataset 3.2 Training 3.3 Data Augmentation 4 Conclusion References Traffic Analysis on Videos Using Deep Learning Techniques 1 Introduction 2 Related Work 3 Proposed System 3.1 Background Subtraction 3.2 Detection and Classification of Vehicles 3.3 Pre-Trained Models 4 Results 5 Future Scope 6 Conclusion References Computer Vision with the Internet of Things (IoT) 1 Introduction 2 What is Computer Vision? 3 Computer Vision Gets Smart 3.1 Early Applications 3.2 Deep Learning Breakthrough 4 Artificial Intelligence, Computer Vision, and Technologies Can be Used to Make Internet of Things 4.1 Four Ways Enterprises are Benefitting from AI and Computer Vision 4.2 Advanced Technologies in IoT Solutions 4.3 Evolving Toward IoT and AI 4.4 Solving the Connectivity Challenge 5 Machine Vision and the Cloud 6 Conclusion References Single Under-Water Image Enhancement Using the Modified Transmission Map and Background Light Estimation 1 1. Introduction 2 Background Study 2.1 The UIFM 2.2 Transmission Map Estimation 3 The proposed U-TMBL 4 Evaluation and Analysis 4.1 Evaluation Metric 4.2 Testing Dataset 4.3 Visual Analysis 4.4 Quantitative Analysis 5 Conclusion References Quality Assessment of the Experimental Data of Wood Structure Using Kanpur Theorem 1 Introduction 1.1 Theoretical Background of CT 2 Material and Methods 2.1 Image Reconstruction from Projections 3 Results and Discussions 3.1 First Kanpur Error Theorem (KT-1) 3.2 Palash 3.3 Rosewood 4 Conclusion References Software Fault Prediction Using Deep Neural Networks 1 Introduction 2 Related Work 3 Proposed Method 3.1 Deep Neural Networks (DNN) 3.2 Dataset 3.3 Performance Measures 4 Experimental Setup and Result Analysis 5 Conclusion References A Hybrid and Multi-objective Approach for Data Leak and Tamper Detection in Healthcare Cloud Data 1 Introduction 2 Related Works 3 Problem Definition 4 Proposed Methodology 5 Experimental Results and Discussion 5.1 Objective I 5.2 Objective II 5.3 Watermarked Image Quality 6 Conclusion and Recommendation References Information Retrieval Using Data Mining 1 Introduction 2 Approved Assessment Procedure 3 A Range of Problems in the Area 4 Issue-Wise Discussion 5 Issue-Wise Solution Assessment Used 6 Solution Approaches for the Specific Issues 6.1 Common Findings 7 Conclusion 8 Future Work References A Novel Approach Toward Detection of Glaucoma Using Machine Learning Techniques 1 Introduction 2 Literature Review 3 Methodology 4 Experimental Results 4.1 Confusion Matrices and Metrics 5 Future Scope 6 Conclusion References Corn Leaf Disease Detection Using ResNext50, ResNext101, and Inception V3 Deep Neural Networks 1 Introduction 2 Background and Literature Survey 3 Proposed Work 4 Experimental Setup and Results Discussion 5 Conclusion and Future Work References SPEG—Semiotics-Based Panel Extraction from Graphic Novel 1 Introduction 2 Related Work 2.1 Character Detection 2.2 Graphic Novel Analysis 2.3 Panel Extraction 3 Proposed Architecture—SPEGYOLO 3.1 Dataset Annotation 3.2 Model Training 3.3 Semiotic Character Detection 3.4 Panel Extraction 3.5 Constraint-Based Extracted Panels 4 Implementation Results 4.1 Experimentation 4.2 Experimental Results 5 Evaluation Metric and Performance Analysis 6 Conclusion and Future Work References Study of Various Word Vectors for Sentiment Analysis 1 Introduction 2 Related Work 3 Proposed Approach 3.1 Methodology 4 Deep Learning Algorithm 4.1 Long Short-Term Memory (LSTM) Network 5 Experiment and Result 6 Conclusion References A Comparative Analysis of 2D Ear Recognition for Constrained and Unconstrained Dataset Using Deep Learning Approach 1 Introduction 2 Methodology 2.1 Database Description and Experimental Setup 3 Result and Performance Evaluation 4 Conclusion References Robust Sliding Mode Controller and Shunt Active Power Filter for Improvement of Power Quality Indices of an Autonomous Microgrid 1 Introduction 2 Literature Review 3 System Modeling 3.1 PV Panel 3.2 Fuel Cell 3.3 Battery 3.4 Shunt Active Power Filter (SAPF) 4 Controller 4.1 PI Controller 4.2 Sliding Mode Controller (SMC) 5 Results and Discussion 6 Conclusion References An Approach for Waste Classification Using Data Augmentation and Transfer Learning Models 1 Introduction 2 Related Work 3 Methodology 3.1 Datasets 3.2 Data Preprocessing and Augmentation 3.3 Synthetic Image Generation Using GANs 3.4 Custom Transfer Learning Models 4 Experimental Results and Discussion 5 Conclusion and Future Work References Adaptive Ridge Edge Detection in Multitemporal Satellite Images 1 Introduction 2 Related Work 3 Proposed System and Implementation 3.1 Data Acquisition 3.2 Gray Level Transformation 3.3 Edge Detection 4 Experimental Results 5 Conclusion References FCM-RGM: Segmentation of Nuclei via Exact Contour Enhancement in Pap Smear Images 1 Introduction 2 Related Works 3 Motivations and Contributions 3.1 Cervical Cancer Dataset Description 4 Preprocessing 5 Median Filtering 5.1 Algorithm 6 Image Enhancement 7 Fuzzy Growing Model 7.1 Region Growing Method 8 Proposed FCM-RGM-Based Abnormality Segmentation 9 Results 10 Segmentation, Review, and Discussion 11 Conclusion References Pulmonary Lung Cancer Classification Using Deep Neural Networks 1 Introduction 2 Related Work 3 Methodology 3.1 Dataset 3.2 Transfer Learning 3.3 Training and Classification 4 Experimental Results 5 Conclusion References Radar Signal Processing for Shaft Rotation Monitoring 1 Introduction 2 Theoretical Background 2.1 Radar Basics 2.2 AWR1642BOOST Evaluation Board 3 Experiment 4 Approach to Signal Processing 4.1 Experiment Result Analysis 4.2 The Proposed Approach for Shaft Rotation Evenness Estimation 5 Conclusion References Machine Learning Assisted MPU6050-Based Road Anomaly Detection 1 Introduction 2 Methodology 2.1 Flowchart Explanation 2.2 Requirement Specification: 2.3 Data Collection 3 Results and Implementation 3.1 Training the ML Model 3.2 Graphical Representation 4 Conclusion and Future Scope References Computer-Aided Leaf Disease Classification Based on Logistic Regression and Transfer Learning Approach 1 Introduction 2 Related Work 3 Materials and Methods 3.1 Dataset Description 3.2 Adaptive Moment Estimation 3.3 Methodology 3.4 Performance Evaluation 4 Result 5 Conclusion References Heart Disease Prediction and Classification Using Machine Learning Models 1 Introduction 2 Proposed Methodology and Algorithm Design 2.1 Algorithm Design for Proposed Model 3 Results and Discussion 4 Conclusion References Artificial Intelligence and Machine Learning Techniques for Analysis of Yoga Pose 1 Introduction 2 Literature Work 3 Methodology 3.1 Dataset Description and Pre-processing 3.2 Convolutional Neural Network 3.3 Machine Learning Classifier 3.4 Parameters Used 4 Experiments Conducted and Results 4.1 Dataset Images 4.2 Creating Skeleton (Tf-Pose Estimation) 4.3 Feature Extraction 4.4 Classification 5 Conclusion References 2D Ear Recognition Using Data Augmentation and Deep CNN 1 Introduction 2 Related Work 3 Material and Methodology 3.1 Dataset Descriptions 3.2 Data Augmentation 3.3 The Proposed ER System 4 Experimental Result and Analysis 5 Conclusion References Binary Pattern for Copy-Move Image Forgery Detection 1 Introduction 2 Related Works 3 Proposed Methodology 3.1 Preprocessing by Wiener Filter 3.2 Feature Extraction by Gabor Filter and CSLBP 3.3 Gabor Filter 3.4 Classification by HNN-DT 3.5 Decision Tree 3.6 Neural Network 4 Experimental Results 4.1 Description of Dataset 4.2 Performance Evaluation 5 Conclusion References Dynamic Deployment Approach to Maximizing Probabilistic Area Coverage in Randomly Scattered WSNs 1 Introduction 2 Probabilistic Area Coverage 3 Proposed Scheme 4 Simulation Results and Discussion 5 Conclusion References Artificial Neural Network Based Image Fusion for Surveillance Application 1 Introduction 2 Literature Review 3 Proposed Methodology 4 Experimental Results 5 Conclusions References Experimental Result Analysis of Extreme Learning Machine with Various Activation Functions: An Application in Credit Scoring 1 Introduction 2 Related Work 3 Results and Discussion 4 Conclusion References Automatic Detection and Classification of Healthy and Unhealthy Plant Leaves 1 Introduction 2 Literature Review 3 Proposed Model 3.1 Machine Learning 3.2 Decision Trees 3.3 Naive Bayes 3.4 Convolutional Neural Network 4 Evaluation of Result 5 Conclusion References Assistive Technology for Blind and Deaf People: A Case Study 1 Introduction 2 Various Technologies to Assist Blind and Deaf People 3 Various Technologies Available in the Market to Assist Blind and Deaf People 4 Challenges and Future Research Scope 5 Conclusion References Trustworthiness of Customer Reviews of E-Commerce Portal Using Supervised and Semi-supervised Machine Learning Techniques 1 Introduction 2 Related Works 3 Proposed Method 4 Experimental Results 5 Conclusions and Future Work References Computer Graphic and Photographic Images Classification—A Transfer Learning Approach 1 Introduction 2 Literature Review 3 Proposed Methodology 3.1 GoogleNet 3.2 Convolution Layer 3.3 Pooling 3.4 Inception 3.5 Fully Connected Layer 4 Experimental Results 5 Conclusion References Hybrid Methods for Increasing Security of IoT and Cloud Data 1 Introduction 1.1 Infrastructure-as-a-Service (IaaS) 1.2 Platform-as-a-Service (PaaS) 1.3 Software-as-a-Service (SaaS) 1.4 Organization of the Paper 2 Related Work 2.1 Implementation of Text Encryption Using Elliptic Curve Cryptography 2.2 Data Security in Cloud Computing Using Various Cryptography Algorithms 2.3 Design and Implementation of Hybrid Encryption Algorithm 2.4 Evaluating the Performance of Symmetric Encryption Algorithms 2.5 Hybrid AES-ECC Model for the Security of Data Over Cloud Storage 2.6 Dual Encryption Model for Preserving Privacy in Cloud Computing 3 Various Encryption Algorithms Used 3.1 Blowfish 3.2 AES Algorithm 3.3 Elliptic Curve Cryptography 4 Methodology and Proposed Hybrid Approach 4.1 Pre-processing the Multimedia Data 4.2 Key Generation and Exchange 4.3 Encryption of Multimedia Data 4.4 Decryption of Multimedia Data 4.5 Post-processing on Decrypted Data 5 Implementation Results 6 Conclusion References Abstractive Text Summarization with Fine-Tuned Transformer 1 Introduction 2 Related Work 3 Approach 3.1 OOV Generation 4 Experimental Results 4.1 Datasets Description 4.2 Results 5 Conclusion References Music Generation Using Deep Learning 1 Introduction 2 Literature Review 3 Dataset 4 Proposed Methodology 4.1 LSTM RNN 4.2 Bidirectional LSTMs RNN 5 Results and Analysis 6 Conclusion References Breast Cancer Classification Using ML on WDBC 1 Introduction 2 Related Works 3 Proposed Model 3.1 Existing System 3.2 Proposed System 4 Methodology 4.1 Dataset 4.2 Dataset Preprocessing 4.3 Machine Learning Algorithms 5 Results and Discussion 6 Conclusion References Kappa: A Measure of Classifier Goodness for Diabetic Retinopathy Using EfficientNet 1 Introduction 2 Related Works 2.1 Diabetic Retinopathy 2.2 Cohen Kappa 2.3 Contrast Enhancement 2.4 Convolution Neural Networks 3 Methodology 4 Results and Discussion 4.1 Dataset 4.2 EDA (Exploratory Data Analysis) 5 Conclusion References Classification of Lung Nodule from CT and PET/CT Images Using Artificial Neural Network 1 Introduction 2 Literature Review 3 Objective 4 Dataset and Method 4.1 Dataset 4.2 Proposed Methodology 5 Results and Discussion 6 Conclusion References An Efficient and Secure Cluster-Based Cooperative Data Transmission for Wireless Ad Hoc Networks IOT Environment 1 Introduction 2 Related Work 3 Contribution of the Paper 4 Methods 4.1 Multi-Hop Cluster-Based Optimal Cooperative MAC (MCOC-MAC) Protocol Transmission Procedure: 4.2 Blockchains-Based Trust Management System in the Ad Hoc IoT Environment 4.3 System Evaluation Metrics 4.4 Metrics for Performance Evolution of the Model 5 Expected Outcome References