Significant difficulties have been posed to healthcare systems all across the world by the COVID-19 epidemic. Due to the high transmission rate of the virus, limited testing resources, and the requirement for speedy and precise diagnosis, COVID-19 patient diagnosis has become a serious problem. CT scans have been a reliable method of diagnosing COVID-19. To aid in the diagnosis of COVID-19, we investigated the SARS-CoV-2 CT-scan dataset in this work. To determine if a CT scan was positive or negative for COVID-19, we turned to a convolution neural network (CNN). We gray scaled and resized the photos to 256x256 before using them in the dataset. In a ratio of 80:20, we divided the dataset into training and validation sets. We used Adam as our optimizer and a learning rate of 0.001 throughout 100 iterations to train the model. On the training set, our accuracy was 93.45%, while on the validation set it was 78.31%. Several criteria, including accuracy, precision, recall, and F1-score, were used to assess the model's efficacy. Overall, the model was 83% accurate on the test data, with precision of 0.87 and recall of 0.81. The confusion matrix demonstrated that the model tended to appropriately categorize COVID-19-negative situations. When tested further, it was discovered that chest CT scans yielded the best results for the model. The study includes a few flaws, such as a small dataset, a homogeneous patient sample, and unstandardized CT scan acquisition procedures. Using larger and more varied datasets and creating standardized techniques for CT scan acquisition are two ways that future research might overcome these restrictions.

Covid-19, Deep Learning, CT scan, CNN;