Dtecht Covid

We train a deep learning algorithm to flag potential covid-19 infected in chest x-rays. The deep learning algorithm used is a Convolutional Neural Network that is 121 layers deep. Due to the lack of a large open-source of covid-19 infected x-ray images, we combine data from five different sources. Combined, the dataset has 17,194 images that are used for training procedure. The model classifies a given chest X- ray image as either a "Normal", "Covid-19", or a "Pneumonia" infection. The trained model has a 0.93 F1 Score and 93.496% accuracy.

Read more about it over here: https://drive.google.com/file/d/1Aeg2BlDlGV6lAOEYbcMCIyAMDd4Nubt3/view?usp=sharing

IEEE Link: https://ieeexplore.ieee.org/document/9824121

Demo Link: https://dtechtcovid-lo3byqgj3a-uc.a.run.app