A Model based on Deep Learning for COVID-19 X-rays Classification

Artificial intelligence in the healthcare sector: comparison of deep learning networks using chest X-ray images

“…Therefore, the results have shown that artificial intelligence (that is, deep learning, which is its subset) is a usable tool in the diagnosis of these diseases. The findings are in parallel with the findings of other studies in the literature ( 22 , 25 , 26 , 51 , 57 ), which suggest that deep learning networks are successful in diagnosing diseases such as COVID-19 and Viral Pneumonia. Therefore, the first of the research questions, “Can deep learning networks (ResNet101, AlexNet, GoogLeNet and Xception) be successful in detecting COVID, Viral Pneumonia and healthy X-ray images?” was answered positively.…”

“…Therefore, the results have shown that artificial intelligence (that is, deep learning, which is its subset) is a usable tool in the diagnosis of these diseases. The findings are in parallel with the findings of other studies in the literature (22,25,26,51,57) (22,25,58) for the detection of COVID-19, but no comparison has been made for the detection of different diseases using CNNs. In some studies [e.g., (28)], there has been a call to examine lung diseases such as pneumonia, lung cancer and COVID-19 together for future studies.…”

“…Apostolopoulos and Mpesiana ( 26 ) have suggested that deep learning models (VGG19, MobileNetv2, Inception, Xception and Inception-ResNetv2) can be used successfully for the detection of this disease in their studies using X-ray images of COVID-19 cases (common bacterial pneumonia, COVID-19, and normal cases). Abd El-Latif and Khalifa ( 25 ) examined deep learning (Alexnet, Resnet18, VGG16, and VGG19) and machine learning [Support Vector Machine (SVM), Decision Trees, and Ensemble algorithm] models in COVID-19 x-rays classification. The findings revealed that VGG19 and SVM integration was more successful (98.61%).…”

“…However, the success rates of different deep learning networks for each disease were not stated. Although different networks of deep learning have been used for COVID-19 detection ( 22 , 25 , 26 ), there are no studies comparing the success levels between networks for multiple diseases such as COVID-19 and pneumonia. Although these networks of deep learning make successful high-level implications in the relevant field by learning from raw data (for example, images), the processes and approaches to classification differ.…”

“…Secondly, although there have been studies in the literature addressing the success rate of deep learning networks in detecting COVID-19 (22,(25)(26)(27), the number of these studies has remained limited. In addition, the success rates of deep learning networks were not clearly compared in these studies, so the success comparison of the networks in image processing was not determined.…”