COVID-19 Detection on X-Ray Images using a Combining Mechanism of Pre-trained CNNs

The COVID-19 infection was sparked by the severe acute respiratory syndrome SARS-CoV-2, as mentioned by the World Health Organization, and originated in Wuhan, Republic of China, eventually extending to every nation worldwide in 2020. This research aims to establish an efficient Medical Diagnosis Support System (MDSS) for recognizing COVID-19 in chest radiography with X-ray data. To build an ever more efficient classifier, this MDSS employs the concatenation mechanism to merge pretrained convolutional neural networks (CNNs) predicated on Transfer Learning (TL) classifiers. In the feature extraction phase, this proposed classifier employs a parallel deep feature extraction approach based on Deep Learning (DL). As a result, this approach increases the accuracy of our proposed model, thus identifying COVID-19 cases with higher accuracy. The suggested concatenation classifier was trained and validated using a Chest Radiography image database with four categories: COVID-19, Normal, Pneumonia, and Tuberculosis during this research. Furthermore, we integrated four separate public X-Ray imaging datasets to construct this dataset. In contrast, our mentioned concatenation classifier achieved 99.66% accuracy and 99.48% sensitivity respectively © 2022. International Journal of Advanced Computer Science and Applications.All Rights Reserved.

Screening Medical Face Mask for Coronavirus Prevention using Deep Learning and AutoML

In the last two years, the COVID-19 pandemic causes a global health crisis around the world. On the other hand, given the current shortage and limits of medical resources, the World Health Organization (WHO) suggests several measures to control the infection rate and avoid depleting limited medical resources. In fact, wearing a medical mask is one of the non-pharmaceutical measures that can be used to limit the spread of this pandemic. This paper aims to present a new deep learning model based on AutoML for medical face mask detection. This proposed model was trained on a publicly available dataset that contained three classes: With mask, Incorrect mask, and Without mask. The achieved results show that the proposed model reaches an Accuracy and sensitivity of 99.74% and 99% respectively. © 2022 IEEE.

Performance Evaluation of Transfer Learning Technique for Automatic Detection of Patients with COVID-19 on X-Ray Images

A new pandemic of coronavirus (COVID19) reported for the first time in Wuhan, China. This new virus has spread rapidly around the world with fever, cough, and difficulty breathing symptoms. In this paper, we propose a Deep Learning based system for the diagnosis of COVID19 disease. This system is based on Transfer Learning technique of six pretrained models. The X-Ray image dataset used contains 2905 images with a resolution of 1024*1024 pixels. A series of preprocessing operations has been applied to this dataset. The performance results obtained in this study confirm that the classification obtained by the Xception network is the most precise for detecting cases infected with COVID19. Our system has achieved accuracy and sensitivity of 98% and 100% respectively. © 2020 IEEE.

Optimization of machine learning algorithms hyper-parameters for improving the prediction of patients infected with COVID-19

In the modeling domain, the selection of appropriate hyper-parameters for classification or prediction algorithms is a difficult task, which has an impact on generalization capacity and classifier performance. In this paper, we compared the performance of five Machine Learning (ML) algorithms from different categories namely: SVM, AdaBoost, Random Forest, XGBoost and Decision Tree. In the first experiment, we adopt a default setting of each model for training and testing. In the second experiment, we use the GridSearch function to find an optimal configuration of the model. The experiments are performed on dataset of anonymous patients with or without COVID-19 disease. The used dataset is obtained from the Albert Einstein Hospital in Sao Paulo, Brazil. To evaluate the reached results, we used different performance evaluation metrics such as: accuracy, precision, recall, AUC and F1-score. The results of the proposed approach have shown that the optimization of the hyper-parameters of the studied learning models leads to an improvement of 18% in terms of Recall. © 2020 IEEE.