The problem CovidCare solves

This app was developed while thinking for solutions like:

1.) This application will be able to track the quarantined as well as non-affected person's and their symptoms will be monitored through our application continuously.
2.) We have provided machine learning models that will help doctors to further confirm if a patient has a possibility or not just by clicking a picture of the person's CT Scan or X-RAY.

Do try both the apps Web App and Android App both are fully functional.

Some Technical terms introduced:

-> Infection Variable (I.V):
It is a custom developed algorithm which will decide if the person is in either of these three states:
Safe => Green Marker
Suspicious => Yellow Marker
Infected => Red Marker
When a user opens the application for the first time he has to fill in a form which will define his IV values and accordingly he/she will be allotted a state mentioned above.

-> Anonymous Red Zones: These are the spots that are generated on our map when a person is detected as infectious.
These spots are created on the map to warn all the users about highly infectious areas.

-> Neighbourhood: This is a map which shows all of the friend's location also their distance from the user and the state(IV) of the friends. All the users are shown in various color markers according to their respective infection variables. This map also shows in real-time the user's location and all the red zones.

Challenges we ran into

While training models we had very little data set for X-ray images of Covid-19 patients, so the accuracy of our model for the X-ray model is lower than our CT-Scan model. We tested with 3-4 different models before we started to the best accuracy for our models.

Accuracy of Models => X-Ray : 85% and CT Scan: 93%

For the friend system development, a good and robust algorithm was quite a hard task to find and after implementing again to find all sorts of bugs.

To keep the infected users anonymous ( i.e while creation of the Red Zone / High IV areas ) was another custom made algorithm that took enormous energy & patience.

Manual configuration of Sqlite database for transferring data between application and server was another tough task but our team aced it ?.