Covid-19 Prediction from Chest X-Ray and CT scan Using Deep Learning Methods
Aarthy. R1, Kasiraja Gunasekarn2, Danish Maqbool Sheikh3, Mallipeddi Jai Sivasankar4, Balaji Selvarajan5
1Assistant Professor, Department of Computer Science and Engineering, Dhanalakshmi Srinivasan Engineering College, Perambalur
2Student, Department of Computer Science and Engineering, Dhanalakshmi Srinivasan Engineering College, Perambalur
3Student, Department of Computer Science and Engineering, Dhanalakshmi Srinivasan Engineering College, Perambalur
4Student, Department of Computer Science and Engineering, Dhanalakshmi Srinivasan Engineering College, Perambalur
5Student, Department of Computer Science and Engineering, Dhanalakshmi Srinivasan Engineering College, Perambalur
ABSTRACT
The novel coronavirus disease of humans, COVID-19, is now thought to be the deadliest sickness it has ever produced. A limited number of COVID-19 test kits are accessible in hospitals because of a shortage of radiologists, and this is also accompanied by a shortage of equipment due to the daily increase in cases as a result of an increase in the number of COVID-19-infected individuals. To Overcome this Several convolutional neural networks (CNN)-based techniques for computer-aided COVID-19 detection based on lung computed tomography (CT) scans and Chest X-ray Images (C-X-rays) have been developed as of late. In this proposed model we used a hybrid dataset of Chest X-ray and CT scans in a single dataset to train the model and predict the COVID-19 cases with both the Chest X-ray and CT scan. However, the presentations of COVID-19 in CT scans and Chest X-Rays are classified into two classes Covid 19 and Normal. Chest X-rays and CT scans provide a different method for detecting Coronavirus early in the disease phase. Using VGG16 and CNN deep learning algorithms, the model identified characteristics from Chest X-ray images and CT scans and categorized them into two groups: Normal and COVID-19. The model had trained with 1600 pictures of two folders test and validation with 2 classes of each folder to test their realism in practical settings. The traditional CNN model has less accuracy in detecting COVID-19 cases, for the betterment of prediction it needs the improved CNN model. The proposed model has the highest accuracy of 94% when compared to the traditional CNN model.
Keywords: CNN, VGG16, COVID-19, Chest X-ray, CT scan, Deep Learning.