Abstract:

In this paper, we utilized deep convolutional networks to classify X-ray pictures into Three groups: Normal, pneumonia, and COVID-19. Our dataset consisted of 575 pictures of COVID-19 patients, 4273 pictures of pneumonia cases, and 1583 Normal pictures. We employed various methods to optimize the network's performance, and we proposed a neural network that combined the VGG16 and ResNet50 networks. This hybrid network obtained the highest accuracy by leveraging the features extracted by the two robust networks. Our network was evaluated using 1288 pictures. and found that it had a Training Accuracy of 98.89% and an overall validation accuracy of 94.5% across all classes.

Description:

Chest X-ray images, COVID-19, Pneumonia, ResNet50 & VGG16 Networks