This is a COVID19 classifier that will predict infected patients from non-infected patients based on a dataset with 145 images on each label.
It's based on a tutorial from pyimagesearch, using an updated dataset with a larger number of images and a ResNet network instead to achieve higher accuracy.
The accuracy, sensitivity, and specificity are the following:
accuracy: 0.9726
sensitivity: 0.9394
specificity: 1.0000
This is for learning and demonstration purposes only. This model is not fitted for any medical applications and should not be used in such enviroment.