Investigation of the Effect of Edge Detection Algorithms in the Detection of Covid-19 Patients with Convolutional Neural Network-Based Features on Chest X-Ray Images

Early diagnosis of COVID-19 is essential to ensure that treatment can be initiated early and to prevent the disease from spreading to other people. In this paper, a deep learningbased method that uses chest X-ray images from normal, COVID-19 and viral pneumonia patients is proposed to enable automatic detection of COVID-19 patients. In addition, Canny, Roberts, Sobel edge detection methods were applied to the images to determine the lesioned area or the perimeter of the area where they are restricted to examine the effect of deep learning on the classification performance. According to the obtained results, when the created deep learning-based model is used in the original data, the classification performance is 94.44% and the highest is 82.30% when edge detection algorithms are used. In addition, although the Sobel algorithm provides better results than other edge detection methods, it can be seen that the classification performance obtained with the original images is higher.