covid19-detection

This code is for predicting COVID-19 from chest Xrays. The model currently has 95% accuracy.The python notebook attached contains the code for training and testing.

Dataset

The dataset was collected from many sources. Currently this git repo is maintaining the COVID-19 Xray dataset: https://github.com/ieee8023/covid-chestxray-dataset

Dataset was divided into 2 classes: COVID-19 and Other (Pneumonia Viral,Pneumonia bacterial and normal)

Model Statistics and Accuracy

VGG16 and InceptionV3 model were utilized as base models (InceptionV3 was finally used), the layers were freezed and on top of it some more layers were added. The model was trained on 318 samples of data and tested on 48 samples. The testing accuracy currently reported is 95%.

ROC Curve

Classification Report: F1 Score, Precision, Recall

Confusion Matrix

[20 1]

[1 26]

Here is the the result of model on one of the samples: