CoviNet: Automated COVID-19 Detection from X-rays using Deep Learning Techniques
Samira Lafraxo, Mohamed El Ansari
Abstract
The novel Coronavirus (COVID19) is an infectious epidemic declared in March 2020 as Pandemic. Because of its easy and rapid transmission, Coronavirus has caused thousands of deaths around the world. Thus, developing new systems for accurate and fast COVID19 detection is becoming crucial. X-ray imaging is used by radiology doctors for the diagnosis of coron-avirus. However, this process requires considerable time. Therefore, artificial intelligence systems can help in reducing pressure on health care systems. In this paper, we propose CoviNet a deep learning network to automatically detect COVID19 presence in chest X-ray images. The suggested architecture is based on an adaptive median filter, histogram equalization, and a convolutional neural network. It is trained end-to-end on a publicly available dataset. Our model achieved an accuracy of 98.62% for binary classification and 95.77% for multi-class classification. As the early diagnosis may limit the spread of the virus, this framework can be used to assist radiologists in the initial diagnosis of COVID19.