A New Convolution Neural Layer Based on Weights Constraints

With the rapid development of artificial intelligence techniques, emerging deep neural networks (DNN) is one of the most effective ways to solve many challenges. Convolution neural networks (CNNs) are considered one of the most popular AI techniques used to extract and analyze meaningful features for image datasets, especially in the medical diagnosis field. In this paper, a proposed constrained convolution layer (COCL) for the CNN model is proposed. The new layer uses a constrained number of weights in each kernel trained in the phase of learning and excludes the others weights with zero values. The proposed method is introduced to extract a special type of feature considering the local shape of a sub-image (window) and topological relations between group pixels. The features extract according to a random distribution of weights in kernels that are determined considering a particular desired percentage. Furthermore, this paper proposed a CNN model architecture that uses COCL rather than the traditional CNN layer (TCL). The efficiency of the method is evaluated using three types of medical image datasets compared with the traditional convolution layer, pre-trained deep neural networks (pre-DNNs), and state-of-art methods. The proposed model outperforms other methods in terms of accuracy and F1 score metrics and exceeds more than 98%, 89%, and 93% for the three datasets used in the evaluation, respectively. © 2022 IEEE.

An Effective Technique using MobileNet-V2 for Face Mask Detection

Recently, the widespread and extremely fatal disease known as the coronavirus spread throughout the entire world. China's Wuhan city served as its first hub for its spread. The COVID-19 outbreak has briefly disrupted our daily routines by affecting worldwide trade and travel. Precautions include hand washing, using hand sanitizer, keeping a safe distance, and most importantly wearing a mask. However, putting on a mask that prevents to some extent airborne droplet transmission will be helpful as a precautionary measure in this pandemic. In the near future, many public service providers will ask the customers to wear masks correctly to avail of their services. However, ensuring that everyone wears a face mask is a difficult chore. Many techniques such as Machine Learning, Deep learning models like CNN, RNN, MobileNet etc. are available to solve this problem. This paper presents a simplified approach using MobileNet-V2 for Face Mask Detection. The model is developed by utilizing TensorFlow, Keras, OpenCV, and Scikit-Learn. The face mask detection model's objective is to identify people's faces and determine whether they are wearing masks at the time they are recorded in the image. An alert will sound if there is a desecration on the scene or in public areas. The challenge with the model is to detect the face mask during motion of a person. Precision, recall, F1-score, support, and accuracy are used to evaluate the system's performance and show its practical pertinency. The system operates with a 99.9% F1 score. The currently developed model will be used in conjunction with embedded camera infrastructure which may then be used to a variety of verticals, including schools, universities, public spaces, airport terminals/gates, etc. © 2023 IEEE.

PSTCNN: Explainable COVID-19 diagnosis using PSO-guided self-tuning CNN

Since 2019, the coronavirus disease-19 (COVID-19) has been spreading rapidly worldwide, posing an unignorable threat to the global economy and human health. It is a disease caused by severe acute respiratory syndrome coronavirus 2, a single-stranded RNA virus of the genus Betacoronavirus. This virus is highly infectious and relies on its angiotensin-converting enzyme 2-receptor to enter cells. With the increase in the number of confirmed COVID-19 diagnoses, the difficulty of diagnosis due to the lack of global healthcare resources becomes increasingly apparent. Deep learning-based computer-aided diagnosis models with high generalisability can effectively alleviate this pressure. Hyperparameter tuning is essential in training such models and significantly impacts their final performance and training speed. However, traditional hyperparameter tuning methods are usually time-consuming and unstable. To solve this issue, we introduce Particle Swarm Optimisation to build a PSO-guided Self-Tuning Convolution Neural Network (PSTCNN), allowing the model to tune hyperparameters automatically. Therefore, the proposed approach can reduce human involvement. Also, the optimisation algorithm can select the combination of hyperparameters in a targeted manner, thus stably achieving a solution closer to the global optimum. Experimentally, the PSTCNN can obtain quite excellent results, with a sensitivity of 93.65% ± 1.86%, a specificity of 94.32% ± 2.07%, a precision of 94.30% ± 2.04%, an accuracy of 93.99% ± 1.78%, an F1-score of 93.97% ± 1.78%, Matthews Correlation Coefficient of 87.99% ± 3.56%, and Fowlkes-Mallows Index of 93.97% ± 1.78%. Our experiments demonstrate that compared to traditional methods, hyperparameter tuning of the model using an optimisation algorithm is faster and more effective. © 2023 Centro Regional de Invest. Cientif. y Tecn.. All rights reserved.

Machine Learning Algorithms for COVID-19 Fake News Detection

Fake news corresponds to distributed information which is not true. It becomes popularized during the 2016 U.S. elections. With the spread of COVID-19 and becoming an epidemic, much information is exchanged around the world. A part of this information is fake having a negative impact on mental health and psychological well-being of people. Because of the importance of this issue, we propose in this work applying several machine learning algorithms to detect COVID-19 fake news. We propose, also, several metrics to evaluate those models and to choose the best among them. Compared to the existing works, we use four classes: Fake, Mostly Fake, True and Mostly True. © 2022 IEEE.

Cross-attention PHV: Prediction of human and virus protein-protein interactions using cross-attention-based neural networks.

Viral infections represent a major health concern worldwide. The alarming rate at which SARS-CoV-2 spreads, for example, led to a worldwide pandemic. Viruses incorporate genetic material into the host genome to hijack host cell functions such as the cell cycle and apoptosis. In these viral processes, protein-protein interactions (PPIs) play critical roles. Therefore, the identification of PPIs between humans and viruses is crucial for understanding the infection mechanism and host immune responses to viral infections and for discovering effective drugs. Experimental methods including mass spectrometry-based proteomics and yeast two-hybrid assays are widely used to identify human-virus PPIs, but these experimental methods are time-consuming, expensive, and laborious. To overcome this problem, we developed a novel computational predictor, named cross-attention PHV, by implementing two key technologies of the cross-attention mechanism and a one-dimensional convolutional neural network (1D-CNN). The cross-attention mechanisms were very effective in enhancing prediction and generalization abilities. Application of 1D-CNN to the word2vec-generated feature matrices reduced computational costs, thus extending the allowable length of protein sequences to 9000 amino acid residues. Cross-attention PHV outperformed existing state-of-the-art models using a benchmark dataset and accurately predicted PPIs for unknown viruses. Cross-attention PHV also predicted human-SARS-CoV-2 PPIs with area under the curve values >0.95. The Cross-attention PHV web server and source codes are freely available at https://kurata35.bio.kyutech.ac.jp/Cross-attention_PHV/ and https://github.com/kuratahiroyuki/Cross-Attention_PHV, respectively.

Deep Learning-Based Efficient Detection of COVID-19

COVID-19 is an infectious disease that mostly affects the lungs and can be fatal. Early detection and isolation are critical for an infected person's survival and preventing the disease from spreading further. To this end, RT-PCR (reverse transcriptase-polymerase chain reaction) and RDT (Rapid Diagnostic Test) are standard diagnoses. But RT-PCR tests have a time-consuming process, and results can take anywhere from 6 hours to 1 day. However, RDT is fast but has a high false diagnosis rate. To tackle this problem, deep learning-based CT imaging can be used as a complement to the present tests. This study proposes an effective deep learning-based approach for identifying Common Pneumonia and COVID-19 from chest CT imaging. COVID-19, Normal, and Common Pneumonia samples received F1-scores of 0.9, 0.9, and 0.8. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Satisfaction prediction of online education in COVID-19 situation using data mining techniques: Bangladesh perspective

This research focuses on the education-based online learning platform. Due to the coronavirus disease (COVID-19) epidemic, online education is gaining global popularity. It has shown how successful it is in investigating the quality of online education at the COVID-19 pandemic situation by 799 students from different academic institutions, schools, colleges, and universities. A Google web form has been utilized as the data gathering mechanism for this survey. This paper perused the prediction of online education through data mining and machine learning approaches in an online program. The data was collected through online questionnaires. To predict online education's satisfaction rate, four different types of classifiers are used e.g., logistic regression classifiers, k-nearest neighbors, support vector machine, naive Bayes classifiers. The key purpose of this research is to find out an answer to a question which is, "are the student's satisfied with starting the new online teaching system, or will it be an ambivalent effect for students in the future?". © 2022 Institute of Advanced Engineering and Science. All rights reserved.

Machine Learning-Based Classification of Chest X-Ray MRI Images into Covid-19 Graphic User Interface

To predict the best performance metrics for the diagnostic pathology of Covid-19 based on MRI image feature extractions, primarily studies analysis are required to determine the optimal setting parameters such as the knn nearest-neighbors, the test size, and the random state. In this investigation, the performance metrics that tell us how much better a model is making prediction are presented. The system is implemented and simulated in Anaconda, and its performance is tested on a real dataset that contains six (06) features and two (02) classes. Each class, an abnormal class (a patient having Covid-19), and a normal class (a patient without Covid-19) consists of 343 instances (images), and 234 instances (images), respectively. At constant random state 66, the performance of test measurements obtained from the simulations results under various test sizes [10%~50%] is carried out when the nearest neighbor knn changes from 1 to 20. For quality analysis to examine and validate the proposed technique, based on the performance metrics, the simulation results achieved an average of train accuracy, test accuracy, precision score, sensitivity, F1-score, and specificity in the interval of (100.0±0.0~74.3±0.9)%, (82.9±3.4~71.6±2)%, (82.5±3.5~66.2±2.8)%, (82.0±6.4~60.1±4.1)%, (80.6±2.3~66.3±3.4)%, and (90.0± 2.3~71.8±3.4)%,respectively. The KNN classifier combined with the optimal setting parameters show better performance, predicting the normal and abnormal class labels accurately. Based on these results, we can further improve the accuracy performance in the range of k = [2~7] and the test size in the range of [10%~35%]. With these primarily studies analysis, we have developed a graphic user interface application to perform the diagnostic of pathology on Covid-19 disease that generates the optimal performance metrics. © 2022 IEEE.

A Two-Phase Classifier Model for Predicting the Drug Satisfaction of the Patients Based on Their Sentiments

Since the corona virus has emerged, genuine clinical resources, such as a paucity of experts and healthcare workers, a lack of adequate equipment and medications, and so on, have reached their peak of inaccessibility. Several people have died as a consequence of the medical profession’s concern. Individuals began self-medicating due to a lack of supply, which exacerbated an already precarious health situation. A rise in new ideas for automation is being spurred by machine learning’s recent success in a varied variety of applications. In this paper, we have proposed a two-phase Decision Tree Classifier based on Artificial Neural networks (DTNN). The work is based on the satisfaction of the drugs among patients with the help of their comments as positive or negative polarity. The dataset of drugs used in this paper is Cymablta and Depopovera. The proposed results are compared with the existing methodology of Support Vector Machine Neural Network (SVMNN). The results are shown in graphical and tabular form which shows the efficiency of the proposed methodology. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Tele-robotic recommendation framework using multi-dimensional medical datasets on COVID-19 classification

The development of robotic partners to take care of daily human life has been expanded recently. Mobile robots have spread their presence within the public environment to assist people in a variety of problematic activities. Mobile Robots are developed with the underlying artificial intelligence technology. Adequate training is provided to the mobile robots under the classifications of supervised learning. The interaction of robots is very important to practice everything that is told to the robotic systems from domestic robots to high-risk work environments that threaten the health of the spinal cord, which focuses on robotic support during the COVID-19 epidemic. In the present research work, a mobile agent is trained using Computerized Tomography (CT) scan reports and X-rays under VGG-16 processing standards for classifying covid and non-covid patients. A hybrid model is designed using Deep Learning Network (DNN) and Convolutional Neural Network (CNN). CNN is trained using images collected using a camera and thermal camera with RGB values ranging from 0 to 255. The advantage of the proposed model in training the mobile agent is making use of CT scan and X-ray images and providing recommendations to the victim about the criticality of being affected by covid. In addition to that, the Machine Learning Algorithm like Decision Tree and Random Forest is constructed and achieved a classification accuracy of 95%. The proposed technique has efficiently provided a reliable recommendation system based on ReLu activation. The other evaluation parameters used to estimate the performance of the proposed model are precision, recall, F1-score. The proposed model achieves 0.84 Precision over the inception technique with 0.79 precision. The reason behind the improvement of accuracy in the present work is the filter used to extract the features. © 2022 The Authors.