Presentation: https://slack-files.com/T010R7Y3M98-F011X9N8M7W-98c69226a9

Inspiration

There is a need for accurate detection of COVID-19

Chest X-Rays/ CTs as second test are more reliable

Support the doctors with AI technology to diagnose and predict the progress of the disease

Increases the survival rate of the patient

What it does

COVID-19 tests are not reliable | Analyze lungs directly

Time consuming analysis protocol! | AI as a faster pipeline

Doctors don’t trust AI black-box | Our pipeline has similar steps as doctors do!

Patient’s info not used fully | We use all data for final severity prediction

How we built it

Segmentation of lungs

Highlight of damaged regions

Calculation of damage degree

Prediction of COVID-19 severity using patients’ data and damage degree

Web frontend for the doctor to see a patients X-Rays and their "severity prediction score". An overlay can be activated to show the machine learning analysis.

Challenges we ran into

Creating the machine learning model.

What's next for Predictovid

Pitching it to doctors.

Share this project:

Updates