Detection of COVID-19 in the time of pandemic when routined checkup takes much time. We have tried to work on CT scan images of lungs of both positive and negative patients and tried to classify them using Machine Learning. In the way of doing so we tried various Machine Learning Classification algorithms and the results derived were not satisfactory. So we started using some pre-trained network architectures and some of them gave good results but not as good as we expected. Since we are working on medical data our aim is to minimize the false-negative errors in our prediction i.e. minimizing the cases in which we predicted that patinet is COVID -ve and actually he is +ve.