WIth the events happening around the world due to the global pandemic, I came up with this idea to build a covid detector based on XRAY images. Because SARS-CoV-2 attacks the epithelial cells lining the lungs, it can be detected from posterioranterior view of the lungs. With limited testing kits, this solution can prove to be favorable and also give results in an instant.

What it does

Classify if person has COVID or not and also identifies the regions in lungs it has spread

How I built it

Trained a ML model to classify xray images of  lungs as covid positive or negative and exposed web sevice using flask

Challenges I ran into

Data scarcity

Accomplishments that I'm proud of

Built the end-to-end solution. The output also shows why the ML model predicts it is covid positive or negative. Highlights on Xray images are shown and how it resulted in the prediction.

What I learned

ML training problems with less data

What's next for COVID_DETECTOR

FIne tune the model, use other information also like geography of the person & other symptoms before confirming as a COVID positive

Try It out



flask, jupyter, keras, tensorflow

Devpost Software Identifier