Search

Purpose: The study aims to develop an unsupervised framework using COVGANs to learn better visual representations of COVID-19 from unlabeled X-ray and CT scans. Methods: We trained multiple-layer GANs to develop the COV-GAN framework on unlabeled X-ray and CT scans. We evaluated the quality of the learned representations using t-SNE visualization, K-means, and GMM clustering. The proposed unsupervised method's performance was compared with leading unsupervised methods for COVID-19 classification on X-ray and CT scans. Results: Our method achieved an accuracy of 75.1% on X-ray scans and 75.7% on CT scans, which is at least 13.9% and 12.3% higher than the leading unsupervised methods for COVID-19 classification on X-ray and CT scans, respectively. The t-SNE visualization, K-means, and GMM clustering showed that our method learned better visual representations of COVID-19 from unlabeled data. Conclusions: Our unsupervised framework using COV-GANs can learn better visual representations of COVID-19 from unlabeled X-ray and CT scans. The learned representations can improve the performance of COVID-19 classification. The outcomes show the potential of unsupervised learning methods to overcome the dearth of labelled data in the medical profession, particularly in times of public health crises like the COVID-19 epidemic. [ABSTRACT FROM AUTHOR]

Apple Scab and Apple Rust are the major classes of apple leaf diseases that gravely affect the apple yield. Seeking an automatic, less expensive, fast yet precise method to detect plant diseases is crucial. Traditional approaches to detect plant diseases using computer vision involve complex and labor-intensive methodologies that rely on image enhancement methods and hand-engineered features. The deep convolutional neural network models are highly favourable in performing image classification with many target classes without involving the arduous phase of feature engineering. In this paper, we utilized the Capsule Neural Network (CapsNet) architecture and modified the network structure by adding additional convolution layers to enhance the model's learning capacity to classify the apple diseases into apple rust, apple scab, healthy, and multiple diseases on the same leaf. Model training was performed on a dataset of images that reflected complex growing conditions observed in the real world. The ability of the model to learn was improved by enhancing the images. Experimentation was conducted on the Kaggle Plant Pathology 2020 - FGVC7 dataset. Experimental study demonstrated a recognition accuracy of 91.37% on the test set, with an overall improvement of 3.67% in accuracy when compared to the research work utilizing the same dataset in literature. Therefore, the proposed method effectively achieves Apple foliar leaf disease detection and surpasses existing state-of-the-art techniques applied to the same dataset. "(Dataset Link: https://www.kaggle.com/c/plant-pathology-2020-fgvc7/data)" [ABSTRACT FROM AUTHOR]

Network-based intrusion detection (N-IDS) is an essential system inside an organization in a smart healthcare enterprise system to prevent the system and its networks from network attacks. A survey of the literature shows that in recent days deep learning approaches are employed successfully for N-IDS using network connections. However, finding the right features from a network connection is a daunting task. This work proposes a multidimensional attention-based deep learning approach for N-IDS that extracts the optimal features for intrusion detection using network payload. The proposed approach includes an embedding that transforms every word in the payload into a 100-dimensional feature vector representation and embedding follows deep learning layers such as a convolutional neural network (CNN) and long short-term memory (LSTM) with attention to extracting optimal features for attack classification. Next, the features of CNN and LSTM layers are concatenated and passed into fully connected layers for intrusion detection. The proposed approach showed 99% accuracy on the KISTI enterprise network payload dataset. In addition, the proposed approach showed 98% accuracy and 99% accuracy on network-based datasets such as KDDCup-99, CICIDS-2017, and WSN-DS and UNSW-NB15 respectively. The good experimental results on various network-based datasets suggest that the proposed N-IDS in smart healthcare enterprise systems is robust and generalizable to detect attacks from different network environments. The proposed approach performed better in all the experiments than the other deep learning-based methods. The model showed a 5% accuracy performance improvement compared to the existing study using the KISTI dataset. In addition, the proposed model has shown similar performances on the other intrusion datasets. The proposed approach serves as a network monitoring tool for efficient and accurate detection of attacks inside an organization on a healthcare enterprise network system. [ABSTRACT FROM AUTHOR]

Leather quality inspection is essential in determining the usable area of the material. As leather is a natural substance, surface defects can have a significant impact on its overall quality and reduce its usability. The automatic identification of surface defects in leather holds great importance in the inspection process. This study presents an innovative method called the Fourier Angular Radial Partitioning (FARP) algorithm for extracting features, specifically tailored for the identification of surface defects in leather. A cutting-edge industrial prototype machine vision system is designed with innovative capabilities to acquire high-quality entire leather surface image accurately. The FARP algorithm leverages a combination of spatial and radial distributed invariant feature descriptors obtained from the magnitude of the Fourier Transform. Furthermore, by partitioning the image into multiple sub-regions enables the FARP to extract features to effectively analyze both prominent flaws like cuts, scars and subtle imperfections like pinholes. The performance of the proposed FARP algorithm is compared to Gray Level Co-occurrence method and Spatial domain features. Correlation analysis is conducted on the extracted features from these three methods to identify the optimal feature set. Leather defects are classified using a multinomial logistic regression model and an ensemble classifier approach with random forest. Various measures, including accuracy, specificity, sensitivity, F-score, Mathew Correlation Coefficient, and ROC analysis using Z-test, are employed for a comprehensive evaluation. The experimental results indicate that the random forest and the proposed FARP feature set, achieves a remarkable classification accuracy of 88.67% and a notable area under the ROC curve of 0.875. This intelligent solution, which integrates FARP with the Random Forest classifier, surpasses the performance of manual expert leather defect classification, highlighting its superior effectiveness. [ABSTRACT FROM AUTHOR]

Alzheimer's Disease (AD) is a common neurological brain disorder that causes the brain cells to die and shrink (Atrophy) gradually, resulting in a continuous decline in one's ability to function independently. Early diagnosis increases the possibility of preventing or delaying the advancement of this mental disorder. Magnetic Resonance Imaging (MRI) offers the potential of non-invasive longitudinal monitoring and plays a vital role as a biomarker of the disease progression. Structural Magnetic Resonance Imaging (sMRI) helps to measure Atrophy, which is considered to be the most dependable biomarker to assess the exact stage and severity of the neuro-degenerative aspect of AD pathology. There are five stages associated with AD, which include Normal Control (NC), Early Mild Cognitive Impairment (EMCI), Mild Cognitive Impairment (MCI), Late Mild Cognitive Impairment (LMCI), and Alzheimer's Disease (AD). In this work, we have used the Alzheimer's Disease Neuroimaging Initiative (ADNI2) sMRI image dataset to measure and classify the stage of AD. In recent years, Convolutional Neural Networks (CNNs) are widely used for medical image analysis. This work focuses on applying different Deep Learning algorithms for the multi-class classification of AD MRI images and proposes the best pre-trained model that can accurately predict the patient's stage. It is observed that ResNet-50v2 gives the best accuracy of 91.84% and f1-score of 0.97 for AD class. Visualization techniques such as Grad-CAM and Saliency Map are applied on the model that gave the best accuracy to understand the region of focus in the image which led to predicting its class. [ABSTRACT FROM AUTHOR]

This paper addresses the diagnosis system of cancer disease using a single setting framework. Most of the radiologists and image specialists are identifying the disease in naked eye. When many conventional systems are used to assess or see a patient's disorder condition, it rarely detects the disease all at once in certain situations. Patients are facing difficulties, when the condition of disease is increasing. Thus, this paper focusses the condition of patient seeing the disease image and developed a single setting framework using a convolutional neural network (CNN) architecture with the help of deep learning approaches. The framework contains several deep learning strategies which are used to determine the patient's relevant illness through affected image, such as mass detection using You-Only-Look-Once (YOLO) approach and the crucial aspect of segmentation by full resolution convolutional networks (FrCN). In last the CNN model is considered for classification. This paper is considered to implement our model using breast cancer disease. The different classifiers and cross-validation tests are taken for evaluating validation matrix items. Comparisons of the existing model with the proposed model are made for improving the diagnosis system. For example, the method Inception V3 for accuracy and AUC are 86.77 and 85.89 on MIAS database whereas proposed model got 99.54 and 98.85 on same evaluation items. Our findings show that the proposed diagnostic model outperforms on conventional detection, segmentation, and classification methods. Thus, our diagnosis process worked much better using deep learning and suggested approaches which will help and facilitate the diagnosis of each contaminated region. In each stage of image processing of the infected region, the suggested diagnostics method could support radiologists. [ABSTRACT FROM AUTHOR]

Damage assessment is one of the most important factors in estimating loss after hurricane. The assessment's findings are very important to identify casualties, temporary housing requirements, financial losses, socio-economic effects, etc. In this paper, our goal is to detect the damage using satellite images captured for the natural disasters like hurricane Harvey. In order to assist the experts to speed up the investigation process for providing the best recommendations this research study presents a compact deep learning-based automatic model to assess hurricane damage using satellite images. TRS-Net (Tropical Revolving Storm Network), model uses flood satellite images to identify the damage caused by the hurricane Harvey. The proposed model effectively extracted features from the satellite images by using depth-wise convolution with varying dilation rates. 23,000 satellite Harvey hurricane images of which 15,000 were damage images caused directly or indirectly and the remaining were the images of normal regions. The model included four clustered layers with convolution and max pooling layer enhanced with batch normalization and dropout to attain better accuracy in detecting the damage caused by hurricane Harvey. The dataset was trained and tested with three different optimizers like Adam, RMSProp and SGD where Adam outperformed the others with 98.57% of testing accuracy. Furthermore, based on the loss and accuracy curve, the performance evaluation of the dataset is analyzed. The experimental findings demonstrate that the suggested TRS-Net model achieves the greatest classification accuracy with 0.99, 0.98, 1.0, 0.96 as F1 Score, precision, sensitivity and specificity respectively. Therefore, the present work assist experts' in efficient hurricane damage assessment and overcome the resource and time intensive manual assessment method. [ABSTRACT FROM AUTHOR]

Around the world, lung disease is a prevalent cause of death and illness. In this article, we propose a lung disease detection system for the automated identification of critical lung diseases: 1. Tuberculosis 2. Regular Pneumonia 3. COVID Pneumonia and healthy individuals from chest X-ray (CXR) images. The proposed system incorporates a two-stage classification method using an ensemble of convolutional neural networks (CNNs). We evaluated the performance of the proposed system using various evaluation metrics with a test set containing 200 normal, 100 Tuberculosis, 200 Regular Pneumonia, and 100 COVID Pneumonia (a total of 600) CXR images. The proposed system achieves an average macro precision, recall, and F1-score of 0.99 and an accuracy of 0.99. The performance of the proposed system based on F1-Score (0.99) for three classes, namely Tuberculosis, Pneumonia, and COVID, is better than the following existing classification frameworks: ensemble of K-ResNets for the same three classes (F1-Score: 0.84), modified ResNet50 (F1-Score:0.89), KNN classifier with fractional multi-channel exponent moments (F1-Score: 0.89), VGG-16 (F1-Score: 0.89), and convolutional sparse support estimator network (F1-Score for COVID: 0.78). The proposed system can also determine the confidence associated with the challenging classification of Regular Vs. COVID Pneumonia. We also utilized visualization techniques like Grad-CAM and Saliency maps to analyze the features learned by the CNNs, to create transparency in the prediction. [ABSTRACT FROM AUTHOR]

This paper has emphasized several sounds for Bird Species Recognition based on their vocalization. Although various techniques have been designed with good equipment for identifying different birds' sounds, still it has limitations to classify or recognize different birds' sounds at dissimilar moments of traveling. Thus, this paper presents the model for recognizing the bird's sound in several situations using Spiking Neural Network (SNN). In this model, Permutation Pair Frequency Matrix Coefficients are considered for the features representing the bird sound frames, after the application of pre-processing method on the moment of recordings. These features are applied as input to the three-layered SNN and subsequently to the hidden layer, which performs the work of spiking if there is any change in the feature. The spikes activate any one of the neuron's outputs, that produces a higher probability class. Further, the SNN along with the Distance Fuzzy Co-clustering algorithm technique used on these bird sounds to achieve optimum results. The performance of this model achieved more than 94% accuracy and a computation time of 4.16 ms better than other existing approaches. [ABSTRACT FROM AUTHOR]

A survey of literature shows that transforming the application files into images and employing deep learning-based models for image classification has been considered as one of the significant directions for malware detection and classification. Mainly, convolutional neural networks (CNN)-based models are successfully employed for Android malware detection and classification. This is mainly due to the reason that this type of malware detection and classification approach is platform independent and has the capability to detect metamorphic and polymorphic malware. The Image-based Android malware detection is resilient to both unpacked and packed malware. Following, this work employs various 26 CNN-based pretrained models and the detailed investigation and analysis of experiments are shown on the Image-based Android malware dataset. Each of these models have the capability to extract its own optimal features and these features are distinct to each other. The penultimate layer features of best performed CNN-based pretrained models are extracted and dimensionality of the features were reduced using kernel principal component analysis (KPCA). The reduced features were fused together and passed into a meta-classifier or stacked classifier for classification. This classifier has two levels; in the first level support vector machine (SVM) and random forest (RForest) machine learning classifier were included for prediction and logistic regression in the second level for classification. The four combinations of fused models are DenseNet, ResNet, InceptionResNet, and EfficientNet. EfficientNet-based fused models showed better performances compared to other fused models and non-fused CNN-based pretrained models. Moreover, the EfficientNet-based fused models outperformed the existing approaches for Android malware detection. All the model performances were shown on two different testing datasets and the proposed model has shown the similar performances on both the testing datasets with attaining better performances during training and testing. This indicates that the proposed model is more generalizable, robust, and it can be used as tool that can be deployed in any application play store. [ABSTRACT FROM AUTHOR]

Feature selection methods have been issued in the context of data classification due to redundant and irrelevant features. The above features slow the overall system performance, and wrong decisions are more likely to be made with extensive data sets. Several methods have been used to solve the feature selection problem for classification, but most are specific to be used only for a particular data set. Thus, this paper proposes wide-ranging approaches to solve maximum feature selection problems for data sets. The proposed algorithm analytically chooses the optimal feature for classification by utilizing mutual information (MI) and linear correlation coefficients (LCC). It considers linearly and nonlinearly dependent data features for the same. The proposed feature selection algorithm suggests various features used to build a substantial feature subset for classification, effectively reducing irrelevant features. Three different datasets are used to evaluate the performance of the proposed algorithm with classifiers which requires a higher degree of features to have better accuracy and a lower computational cost. We considered probability value (p value <0.05) for feature selection in experiments on different data sets, then the number of features is selected (such as 7, 5, and 6 features from mobile, heart, and diabetes data set, respectively). Various accuracy is considered with different classifiers; for example, classifier Nearest_Neighbors made accuracy such as 0.92225, 0.88333, 0.86250 for mobile, heart, and diabetes data sets, respectively. The proposed model is adequate as per the evaluation of several real-world data sets. [ABSTRACT FROM AUTHOR]

Abstract Pediatric pneumonia has drawn immense awareness due to the high mortality rates over recent years. The acute respiratory infection caused by bacteria, viruses, or fungi infects the lung region and hinders oxygen transport, making breathing difficult due to inflamed or pus and fluid-filled alveoli. Being non-invasive and painless, chest X-rays are the most common modality for pediatric pneumonia diagnosis. However, the low radiation levels for diagnosis in children make accurate detection challenging. This challenge initiates the need for an unerring computer-aided diagnosis model. Our work proposes Contrast Limited Adaptive Histogram Equalization for image enhancement and a stacking classifier based on the fusion of deep learning-based features for pediatric pneumonia diagnosis. The extracted features from the global average pooling layers of the fine-tuned MobileNet, DenseNet121, DenseNet169, and DenseNet201 are concatenated for the final classification using a stacked ensemble classifier. The stacking classifier uses Support Vector Classifier, Nu-SVC, Logistic Regression, K-Nearest Neighbor, Random Forest Classifier, Gaussian Naïve Bayes, AdaBoost classifier, Bagging Classifier, and Extra-trees Classifier for the first stage, and Nu-SVC as the meta-classifier. The stacking classifier validated using Stratified K-Fold cross-validation achieves an accuracy of 98.62%, precision of 98.99%, recall of 99.53%, F1 score of 99.26%, and an AUC score of 93.17% on the publicly available pediatric pneumonia dataset. We expect this model to greatly help the real-time diagnosis of pediatric pneumonia. [ABSTRACT FROM AUTHOR]

Purpose: According to the literature of International Diabetes Federation reports in 2019, one of the causes of adulthood death is diabetes. Diabetes is a chronic metabolic disease that has a long-term impact on the individual's well-being. Insulin resistance and abnormal glucose metabolism are hallmarks of type 2 diabetes mellitus (T2DM). Noncommunicable Diseases (NCD), such as diabetes, are caused by poor eating habits and way of life. Methods: In this work, several risk prediction models for Type 2 Diabetes Mellitus have been proposed. Using multivariate analysis to assist patients with risk stratification within a population, this study also attempts to determine the relationship between lifestyle behavior patterns and diabetes. Considering that a number of diabetes predictive models proposed by previous researchers rely primarily on specific medical measurement data, but lack external diabetic factors, external diabetic factors are necessary for the development of diabetes predictive models. Glycated Hemoglobin (HbA1c) is utilized in the form to diagnose diabetes due to its efficiency and patient-friendliness. In addition, contrary to the widespread belief that machine learning is superior in many ways, a number of issues, such as racial bias, must be considered when deploying machine learning in the healthcare industry. Because HbA1c is influenced by external factors such as race and ethnicity, the Asia-Pacific region has a variety of HbA1c cut-off points. As a result, restricting the population scope must be viewed as the best approach in this task to permit improved accuracy and assurance. Results: A detailed experimental analysis of various machine learning model performances was evaluated on a standard Type-2 Diabetes dataset. Random forest with SMOTE oversampling and PCA technique outperformed the other methods for Type-2 Diabetes prediction with a recall of 65%, precision of 89%, and f1-score of 75%. The machine learning models performances were evaluated on various test cases during testing. Conclusions: This study leveraging a dataset of medical report and lifestyle behavior factor to predict the HbA1c category. The proposed approach can be deployed as a tool at medical centers to serve as an early Type-2 disease diagnosis tool and in addition, the tool can assist the medical experts in accurately predicting the Type-2 diabetes disease. [ABSTRACT FROM AUTHOR]

Purpose: the problem of big data analytics and health care support systems are analyzed. There exist several techniques in supporting such analytics and robust support systems; still, they suffer to achieve higher performance in disease prediction and generating the analysis.. For a hospital unit, maintaining such massive data becomes a headache. However, still, the big data can be accessed towards analyzing the bio signals obtained from the human body for the detection and prediction of various diseases. To overcome the deficiency, an efficient Health Care Big Data Analytics Model (HCBDA) is presented, which maintains a massive volume of data in the data server. Methods: The HCBDA model monitors the patients for their current state in their cardiac and anatomic conditions to predict the diseases and risks. To perform analysis on health care, the model has accessed the data location by discovering the possible routes to reach the source. The monitored results on blood pressure, temperature, and blood sugar are transferred through the list of routes available. The network is constructed with a list of sensor nodes and Internet of Things (IoT) devices, where the sensor attached to the patient initiates the transmission with the monitored results. The monitored results are transferred through the number of intermediate nodes to the monitoring system, which accesses the big data to generate intelligence. The route selection is performed according to the value of Trusted Forwarding Weight (TFW) and Trusted Carrier Weight (TCW). At each reception, the features from the packet are extracted, and obtained values are fed to the decisive support system. The decisive support system cluster the big data using the FDS clustering algorithm, and the classification is performed by measuring the feature disease class similarity (FDCS). According to the class identified, the method would calculate Disease Prone Weight (DPW) to generate recommendations to the medical practitioner. Results: The unique Health Care Big Data Analytics (HCBDA) paradigm for patient-centered healthcare using wireless sensor networks and IoT devices was described. The patient's bio signals are watched in order to provide medical assistance. In comparison to the previous methods, the proposed approach helps to generate higher performance in disease prediction accuracy up to 96%. Conclusion: The value of Trusted Forwarding Weight (TFW) and Trusted Carrier Weight is used to determine the route (TCW). Sensor based IoT values like Pressure glucose, pulse oximeter, and temperature etc. the following parameters like classification accuracy and false ratio are calculated based on efficient machine learning model. The crucial support system receives the values it receives after each reception together with the features that were derived from the packet. The classification is carried out by calculating the Feature Disease Class Similarity, and the decision support system clusters the huge data using the FDS clustering technique. [ABSTRACT FROM AUTHOR]

Pneumonia is an acute respiratory infection caused by bacteria, viruses, or fungi and has become very common in children ranging from 1 to 5 years of age. Common symptoms of pneumonia include difficulty breathing due to inflamed or pus and fluid-filled alveoli. The United Nations Children's Fund reports nearly 800,000 deaths in children due to pneumonia. Delayed diagnosis and overpriced tests are the prime reason for the high mortality rate, especially in underdeveloped countries. A time and cost-efficient diagnosis tool: Chest X-rays, was thus accepted as the standard diagnostic test for pediatric pneumonia. However, the lower radiation levels for diagnosis in children make the task much more onerous and time-consuming. The mentioned challenges initiate the need for a computer-aided detection model that is instantaneous and accurate. Our work proposes a stacked ensemble learning of deep learning-based features for pediatric pneumonia classification. The extracted features from the global average pooling layer of the fine-tuned Xception model pretrained on ImageNet weights are sent to the Kernel Principal Component Analysis for dimensionality reduction. The dimensionally reduced features are further trained and validated on the stacking classifier. The stacking classifier consists of two stages; the first stage uses the Random-Forest classifier, K-Nearest Neighbors, Logistic Regression, XGB classifier, Support Vector Classifier (SVC), Nu-SVC, and MLP classifier. The second stage operates on Logistic Regression using the first stage predictions for the final classification with Stratified K-fold cross-validation to prevent overfitting. The model was tested on the publicly available pediatric pneumonia dataset, achieving an accuracy of 98.3%, precision of 99.29%, recall of 98.36%, F1-score of 98.83%, and an AUC score of 98.24%. The performance shows its reliability for real-time deployment in assisting radiologists and physicians. [ABSTRACT FROM AUTHOR]

This paper proposes a multichannel deep learning approach for lung disease detection using chest X-rays. The multichannel models used in this work are EfficientNetB0, EfficientNetB1, and EfficientNetB2 pretrained models. The features from EfficientNet models are fused together. Next, the fused features are passed into more than one non-linear fully connected layer. Finally, the features passed into a stacked ensemble learning classifier for lung disease detection. The stacked ensemble learning classifier contains random forest and SVM in the first stage and logistic regression in the second stage for lung disease detection. The performance of the proposed method is studied in detail for more than one lung disease such as pneumonia, Tuberculosis (TB), and COVID-19. The performances of the proposed method for lung disease detection using chest X-rays compared with similar methods with the aim to show that the method is robust and has the capability to achieve better performances. In all the experiments on lung disease, the proposed method showed better performance and outperformed similar lung disease existing methods. This indicates that the proposed method is robust and generalizable on unseen chest X-rays data samples. To ensure that the features learnt by the proposed method is optimal, t-SNE feature visualization was shown on all three lung disease models. Overall, the proposed method has shown 98% detection accuracy for pediatric pneumonia lung disease, 99% detection accuracy for TB lung disease, and 98% detection accuracy for COVID-19 lung disease. The proposed method can be used as a tool for point-of-care diagnosis by healthcare radiologists.Journal instruction requires a city for affiliations; however, this is missing in affiliation 3. Please verify if the provided city is correct and amend if necessary.correct [ABSTRACT FROM AUTHOR]

Purpose: The paper is to study a review of the employment of deep learning (DL) techniques inside the healthcare sector, together with the highlight of the strength and shortcomings of existing methods together with several research ultimatums. Our study lays the foundation for healthcare professionals and government with present-day inclinations in DL-based data analytics for smart healthcare. Methods: A deep learning-based technique is designed to extract sensor displacement effects and predict abnormalities for activity recognition via Artificial Intelligence (AI). The presented technique minimizes the vanishing gradient issue of Recurrent Neural Networks (RNN), thereby reducing the time for detecting abnormalities with consideration of temporal and spatial factors. Proposed Moran Autocorrelation and Regression-based Elman Recurrent Neural Network (MAR-ERNN) introduced. Results: Experimental results show the feasibility of the proposed method. The results show that the proposed method improves accuracy by 95% and reduces execution time by 18%. Conclusion: MAR-ERNN performs well in the activity recognition of health status. Collectively, this IoT-enabled smart healthcare system is utilized by enhancing accuracy, and minimizing time and overhead reduction. [ABSTRACT FROM AUTHOR]

This paper addresses a model to secrete the information of one image under another without losing quality of image. Different approaches have been utilized for image hiding as needed, but multiple images maintain secrecy with information under another image is a challenging task. Thus, the framework is proposed to sustain the secrecy of an original image from another image. The proposed system collects random images through ImageNet and uses them as per the requirements of secrete images. The framework is used the deep neural networks method to build secrete information of multiple images under a single image. The enormous transfer of images is used to select standard image modifications using advanced deep learning approaches. It develops the significance of the critical framework that alleviates the choice of finding the hidden image information. Two vital methods such as Peak Signal to Noise Ratio (PSNR) and the Structural Similarity Index (SSIM) are used to find out difference between host and secret image by their corresponding evaluation scores. It produces the confidentiality of the image with the help of the host image. Therefore, data from several images are protected under a single image. The different image data are experimented with good performance. For comparative analysis, the accuracy is better in retrieving two secrete images on all experiments, like approximate accuracy is 100%. Still, when we considered PSNR and SSIM scores on the same two secrete images, accuracy became less than 50%. [ABSTRACT FROM AUTHOR]

Literature survey shows that the recommendation systems have been largely adapted and evaluated in various domains. Due to low performances from various cyber attacks, the adoption of recommender system is in the initial stage of defense systems. One of the most common attacks for recommender system is shilling attack. There are some existing techniques for identifying the shilling attacks built in the user ratings patterns. The performance of ratings on target items differs between the attack user profiles and actual user profiles. To differentiate the certain profiles, the affected profiles are known as attack profiles. Besides the shilling attacks, real cyber attacks are taking place in the community which are being solved by Petri Net methods. These attacks can be falsely predicted (shilling attacks) by the users which can raise security threats. For identifying various shilling attacks without a priori knowledge, Recommendation System suffers from low accuracy. Basically, recommendation attack is split into nuke and push attack that encourage and discourage the recommended target item. The strength of shilling attack is usually measured by filler size and attack size. An experiment over unsupervised machine learning algorithms with filler size 3% over 3%, 5%, 8% and 10% attack sizes is presented for Netflix dataset. Furthermore, we conducted an experiment on data of 26 K videos on the Trending YouTube Video Statistics, to predict the user preferences for a particular genre of videos using Machine Learning Algorithms. Based on the results, it observed that the Boosted Decision tree performs the best with an accuracy of 99 percent. [ABSTRACT FROM AUTHOR]

This paper addresses the data privacy based on interactive computation using an optimization model in data mining. When data are computed or sharing among users in online, it needs to maintain privacy for all computation during sharing of data. But user choice-based privacy is not available when sharing of data is required for data mining computation which is a big challenge for data privacy. Thus, we proposed the framework for anonymity of data privacy using various methods of multi-objective models as per the requirement of privacy. The proposed framework is designed with the help of two objects such as computational cost and privacy based on optimization model. Our framework maintains the balance between above objects as per user demands, i.e., increasing the privacy with decreasing the computational cost. In this model, the domain of privacy and computational cost for optimization problem solves the entity privacy requirements in a computing environment. We have used various methods such as Gaussian and uniform distribution, confidence interval, activation function, linear membership function with distinguish manner for maintaining of privacy and cost. As per the uniform distribution and parameter α-cut value for noise data, the optimal value is made accordingly. Example: for α = 0.2, and uniform distribution (− 1, 1), the optimal value is 0.0058. Similarly, as per different α values, classifiers result is different like α = 0.2 and 0.4, Multilayer perceptron values are 4.01 and 1.61 respectively. The solution of the proposed model controls the amount of privacy with complete freedom of choice of users with utmost flexibility. [ABSTRACT FROM AUTHOR]

The decision-making process is very crucial in healthcare, which includes quick diagnostic methods to monitor and prevent the COVID-19 pandemic disease from spreading. Computed tomography (CT) is a diagnostic tool used by radiologists to treat COVID patients. COVID x-ray images have inherent texture variations and similarity to other diseases like pneumonia. Manually diagnosing COVID X-ray images is a tedious and challenging process. Extracting the discriminant features and fine-tuning the classifiers using low-resolution images with a limited COVID x-ray dataset is a major challenge in computer aided diagnosis. The present work addresses this issue by proposing and implementing Histogram Oriented Gradient (HOG) features trained with an optimized Random Forest (RF) classifier. The proposed HOG feature extraction method is evaluated with Gray-Level Co-Occurrence Matrix (GLCM) and Hu moments. Results confirm that HOG is found to reflect the local description of edges effectively and provide excellent structural features to discriminate COVID and non-COVID when compared to the other feature extraction techniques. The performance of the RF is compared with other classifiers such as Linear Regression (LR), Linear Discriminant Analysis (LDA), K-nearest neighbor (kNN), Classification and Regression Trees (CART), Random Forest (RF), Support Vector Machine (SVM), and Multi-layer perceptron neural network (MLP). Experimental results show that the highest classification accuracy (99. 73%) is achieved using HOG trained by using the Random Forest (RF) classifier. The proposed work has provided promising results to assist radiologists/physicians in automatic COVID diagnosis using X-ray images. [ABSTRACT FROM AUTHOR]

Purpose: This study proposes to identify potential liver patients based on the results of a liver function test performed during a health screening to search for signs of liver disease. It is critical to detect a liver patient at an early stage in order to treat them effectively. A liver function test's level of specific enzymes and proteins in the blood is evaluated to determine if a patient has liver disease. Methods: According to a review of the literature, general practitioners (GPs) rarely investigate any anomalies in liver function tests to the level indicated by national standards. The authors have used data pre-processing in this work. The collection has 30691 records with 11 attributes. The classification model is utilized to construct an effective prediction system to aid general practitioners in identifying a liver patient using data mining. Results: The collected results indicate that both the Naïve Bayes and C4.5 Decision Tree models give accurate predictions. However, given the C4.5 model offers more accurate predictions than the Naïve Bayes model, it can be assumed that the C4.5 model is superior for this research. Consequently, the liver patient prediction system will be developed using the rules given by the C4.5 Decision Tree model in order to predict the patient class. The training set, suggested data mining with a classification model achieved 99.36% accuracy and on the testing set, 98.40% accuracy. On the training set, the enhanced accuracy relative to the current system was 29.5, while on the test set, it was 28.73. In compared to state-of-the-art models, the proposed approach yields satisfactory outcomes. Conclusion: The proposed technique offers a variety of data visualization and user interface options, and this type of platform can be used as an early diagnosis tool for liver-related disorders in the healthcare sector. This study suggests a machine learning-based technique for predicting liver disease. The framework includes a user interface via which healthcare providers can enter patient information. [ABSTRACT FROM AUTHOR]

Purpose: Breast cancer is a tumour affecting the breast tissues and is detected commonly in women. Computer-aided diagnosis aids in mass screening and early detection of breast cancer. Technological advancements in deep learning are driving the growth of automated breast cancer diagnosis. Various breast cancer image modalities like histopathology, mammograms, thermography, and ultrasound-based images are used for detecting breast cancer. This paper tends to explore a review of current studies related to breast cancer classification using deep learning techniques. Methods: The review highlights the imaging modalities, publicly available datasets, augmentation techniques, preprocessing techniques, transfer learning approaches, and deep learning techniques used by various researchers in the process of breast cancer early detection. In addition, the study also presents a comprehensive review of the performance of the existing deep learning algorithms, challenges, and future research directions. Results: The various methods proposed so far have been compared based on performance metrics including accuracy, sensitivity, specificity, AUC, and F-measure. Many research works have attained an accuracy of more than 90% in the classification and analysis of breast cancer detection using histopathological images. Conclusion: This review has pointed out various limitations that have to be improved for accurately identifying breast cancer using image modalities. Incorrect labelling of data due to observer variations with regard to image segmentation datasets, computation time, and Memory overhead has to be analyzed in future for resulting an enhanced CAD system. [ABSTRACT FROM AUTHOR]

Purpose: Research into predictive analytics, which helps predict future values using historical data, is crucial. In order to foresee future instances of COVID-19, a method based on the Seasonal ARIMA (SARIMA) model is proposed here. Additionally, the suggested model is able to predict tourist arrivals in the tourism business by factoring in COVID-19 during the pandemic. In this paper, we present a model that uses time-series analysis to predict the impact of a pandemic event, in this case the spread of the Coronavirus pandemic (Covid-19). Methods: The proposed approach outperformed the Autoregressive Integrated Moving Average (ARIMA) and Holt Winters models in all experiments for forecasting future values using COVID-19 and tourism datasets, with the lowest mean absolute error (MAE), mean absolute percentage error (MAPE), mean squared error (MSE), and root mean squared error (RMSE). The SARIMA model predicts COVID-19 and tourist arrivals with and without the COVID-19 pandemic with less than 5% MAPE error. Results: The suggested method provides a dashboard that shows COVID-19 and tourism-related information to end users. The suggested tool can be deployed in the healthcare, tourism, and government sectors to monitor the number of COVID-19 cases and determine the correlation between COVID-19 cases and tourism. Conclusion: Management in the tourism industries and stakeholders are expected to benefit from this study in making decisions about whether or not to keep funding a given tourism business. The datasets, codes, and all the experiments are available for further research, and details are included in the appendix. [ABSTRACT FROM AUTHOR]

Intracranial Haemorrhage (ICH) occurring due to any injury to the brain is a fatal condition and its timely diagnosis is critically important. In this work, we propose a complete one-stop model for the identification of Intracranial Haemorrhage (ICH) and for the segmentation of ICH regions in Computerized Tomography (CT) images. The proposed method incorporates Residual Neural Network (ResNet) architecture for ICH identification and further segments the ICH region using an Adversarial Network called SegAN. This work incorporates a data augmentation method using CycleGAN, to solve the problem of class imbalance in the ICH dataset, leading to improved performance in the ICH identification task. CycleGAN is trained to convert a non-ICH CT slice into a synthetic CT slice with ICH, thereby augmenting the ICH sub-class, where there is a lack of data points. The proposed method achieved a macro average F1-score of 0.91 and a specificity of 0.99 and a sensitivity of 0.80 in the ICH identification task. Also, the proposed method works as a segmentation tool for all the five ICH sub-types and achieved a dice score of 0.32 and a mean Intersection Over Union (IOU) of 0.22. Thus, our proposed ICH identification and segmentation model can aid doctors in the accurate and timely diagnosis of ICH. [ABSTRACT FROM AUTHOR]

The epidemic of malaria has caused many deaths for decades around the world. To shorten the morbidity of malaria, accurate and fast diagnostic tools are to be implied with Artificial Intelligence. In this research, an automated, fast, and accurate diagnostic model for one-shot detection and classification of malaria thin blood smears were developed using the Deep Siamese Capsule Network (D-SCN) and contributions for our research are threefold. Firstly, we proposed the D-SCN model which consists of two crucial phases which are feature extraction and feature discrimination. Secondly, we implied an end-to-end trained capsule network with an imperative routing (IR) mechanism for the feature extraction phase to capture feature invariances. Finally, at the feature discrimination phase, Lorentz, L1 and L2 similarity metrics were proposed for the dissimilation of features. During experimentation, it is observed that the Lorentz similarity metric provided more discriminative capability by acquiring the least MSE at most of the instances. Further, an algorithm is proposed to obtain faster convergence by cautiously tuning the hyperparameters, and this aided in decreasing the noise scale in training the D-SCN. The experimental outcomes proved that D-SCN outperformed with the highest detection accuracy of 97.24% for Lorentz as a similarity measure, and the Capsule network with IR mechanism outperformed with the highest classification accuracy of 98.89%. To our knowledge, the proposed research implications are the first applications of D-SCN for one-shot detection and classification of thin blood smears with state-of-the-art performance. [ABSTRACT FROM AUTHOR]

Diagnosing COVID-19, current pandemic disease using Chest X-ray images is widely used to evaluate the lung disorders. As the spread of the disease is enormous many medical camps are being conducted to screen the patients and Chest X-ray is a simple imaging modality to detect presence of lung disorders. Manual lung disorder detection using Chest X-ray by radiologist is a tedious process and may lead to inter and intra-rate errors. Various deep convolution neural network techniques were tested for detecting COVID-19 abnormalities in lungs using Chest X-ray images. This paper proposes deep learning model to classify COVID-19 and normal chest X-ray images. Experiments are carried out for deep feature extraction, fine-tuning of convolutional neural networks (CNN) hyper parameters, and end-to-end training of four variants of the CNN model. The proposed CovMnet provide better classification accuracy of 97.4% for COVID-19 /normal than those reported in the previous studies. The proposed CovMnet model has potential to aid radiologist to monitor COVID-19 disease and proves to be an efficient non-invasive COVID-19 diagnostic tool for lung disorders. [ABSTRACT FROM AUTHOR]

This paper addresses disease analysis using machine learning approaches in healthcare system. Several approaches have been used to identify various disease as their corresponding model, but generic model for detecting disease is a challenging task. Thus, this paper proposed the model for disease detection using machine learning approaches with various methodologies such as support vector machine (SVM), K-nearest neighbours, random forests, artificial neural networks (ANNs), and logistic regression. This paper is also used an evaluation matrix with different parameters for performance analysis. The experimental performance is identified as per proposed model through the evaluation matrix. The outcomes disclose that the ANNs method performed good compare to others based on accuracy (97.94%), precision (96.78%), and F1-score (97.87%), respectively. The correlation approach also determined the number of attributes with very close diagonal values i.e., 100%. The comparative approaches are strongly analysed in experiments for clarity of performance. [ABSTRACT FROM AUTHOR]

Literature survey shows that convolutional neural network (CNN)-based pretrained models have been largely used for CoronaVirus Disease 2019 (COVID-19) classification using chest X-ray (CXR) and computed tomography (CT) datasets. However, most of the methods have used a smaller number of data samples for both CT and CXR datasets for training, validation, and testing. As a result, the model might have shown good performance during testing, but this type of model will not be more effective on unseen COVID-19 data samples. Generalization is an important term to be considered while designing a classifier that can perform well on completely unseen datasets. Here, this work proposes a large-scale learning with stacked ensemble meta-classifier and deep learning-based feature fusion approach for COVID-19 classification. The features from the penultimate layer (global average pooling) of EfficientNet-based pretrained models were extracted and the dimensionality of the extracted features reduced using kernel principal component analysis (PCA). Next, a feature fusion approach was employed to merge the features of various extracted features. Finally, a stacked ensemble meta-classifier-based approach was used for classification. It is a two-stage approach. In the first stage, random forest and support vector machine (SVM) were applied for prediction, then aggregated and fed into the second stage. The second stage includes logistic regression classifier that classifies the data sample of CT and CXR into either COVID-19 or Non-COVID-19. The proposed model was tested using large CT and CXR datasets, which are publicly available. The performance of the proposed model was compared with various existing CNN-based pretrained models. The proposed model outperformed the existing methods and can be used as a tool for point-of-care diagnosis by healthcare professionals. [ABSTRACT FROM AUTHOR]

The Severe Acute Respiratory Syndrome (SARS)-CoV-2 virus caused COVID-19 pandemic has led to various kinds of anxiety and stress in different strata and sections of the society. The aim of this study is to analyse the sleeping and anxiety disorder for a wide distribution of people of different ages and from different strata of life. The study also seeks to investigate the different symptoms and grievances that people suffer from in connection with their sleep patterns and predict the possible relationships and factors in association with outcomes related to COVID-19 pandemic induced stress and issues. A total of 740 participants (51.3% male and 48.7% female) structured with 2 sections, first with general demographic information and second with more targeted questions for each demographic were surveyed. Pittsburgh Sleep Quality Index (PSQI) and General Anxiety Disorder assessment (GAD-7) standard scales were utilized to measure the stress, sleep disorders and anxiety. Experimental results showed positive correlation between PSQI and GAD-7 scores for the participants. After adjusting for age and gender, occupation does not have an effect on sleep quality (PSQI), but it does have an effect on anxiety (GAD-7). Student community in spite of less susceptible to COVID-19 infection found to be highly prone to psychopathy mental health disturbances during the COVID-19 pandemic. The study also highlights the connectivity between lower social status and mental health issues. Random Forest model for college students indicates clearly the stress induced factors as anxiety score, worry about inability to understand concepts taught online, involvement of parents, college hours, worrying about other work load and deadlines for the young students studying in Universities. [ABSTRACT FROM AUTHOR]

A wireless sensor network is a potential technique which is most suitable for continuous monitoring applications where the human intervention is not possible. It employs large number of sensor nodes, which will perform various operations like data gathering, transmission and forwarding. An optimal Q-learning based clustering and load balancing technique using improved K-Means algorithm is proposed. It contains two phases namely clustering phase and node balancing phase. The proposed algorithm uses Q-learning technique for deploying sensor nodes in appropriate clusters and cluster head CH election. In the clustering phase, the node will be placed in appropriate clusters based on the computation of the mean values. Once the sensors are placed in an appropriate cluster, then the cluster will be divided into 'k' partitions. The node which is having maximum residual energy in each partition will be elected as the partition head PH. In node balancing phase, the number of sensors in each partition will be evenly distributed by considering the area of the cluster and the number of sensors inside the cluster. Among the PHs, the node which is having residual energy to the maximum and also having the minimal distance to the sink is elected as the CH. The residual energy of the CH is monitored periodically. If it falls below the threshold level, then another partition head PH which is having residual energy to the maximum level and possessing minimum distance to the sink node will be elected as CH. The proposed Q-Learning based clustering technique maximize the reward by considering the throughput, end-to-end delay, packet delivery ratio and energy consumption. Finally, the performance of the Q-learning based clustering algorithm is evaluated and compared existing k-means based clustering algorithms. Our results indicate that the proposed method reduces end to end delay by 8.23%, throughput is increased by 2.34%, network lifetime is increased by 3.34%, packet delivery ratio is improved by 1.56%. [ABSTRACT FROM AUTHOR]