Use of Deep Learning Models to detect COVID-19 from Chest X-Rays

Kandhari, R.; Negi, M.; Bhatnagar, P.; Mangipudi, P.; Ieee,

The COVID 19 pandemic has spread rapidly across the globe in the last one year and the global response to it especially from the artificial intelligence community has been tremendous. It was seen from chest CT scans and chest X-rays that the respiratory system is hugely affected by the corona virus. In developing countries like India, a rapid and low-cost diagnostic tool is of prime importance for mass screening of population. Given the recent success of deep learning techniques in solving real-time problems, this paper presents application of deep learning models to detect and classify COVID 19 from chest radiographs. Dataset of 2727 images was constructed for this study and analysis from various open source resources like Kaggle and Github. Several pre-trained models have been used for experimentation. Among all the models, VGG-16 model achieved an impressive classification accuracy of 98.9% and F1-score of 0.984 with high sensitivity and specificity as well.