Mixed attention and regularized COVID-19 network: An approach to detection of COVID-19 with chest x-ray images

Coronavirus Disease 2019 (COVID-19) has led to a global pandemic in the year 2020 and the cases are dynamically increasing and active all over the world. COVID-19 is caused due to Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). It is a human-to-human transmissible disease which has severely affected people especially with weaker immunity, and is detected through Reverse Transcription Polymerase Chain Reaction (RT-PCR). RT-PCR is a lethargic process and therefore intelligent systems are proposed which uses chest images for early detection of COVID-19. This paper proposes a regularized and attentive intelligent system called 'Mixed Attention & Regularized COVID-19 Network (MARCOV19-Net)' for detection of COVID-19 using chest X-Ray images. The performance of MARCOV19-Net is compared with VGG-16, Regularized COVID-19 Deep Convolutional Network (RCOV19-DCNet) and Mixed Attention and unregularized COVID-19 Network (MACOV19-Net), and with other state-of-the-art models. MARCOV19-Net has achieved the highest F-score, ROC and AUC of 98.76%, 99.4% and 99.6%, respectively.