PhysicAIn

Inspiration

We got inspired by the level of accuracy that machine learning models can identify objects using images. We want to use these capabilities to benefit patients by giving them accurate results in seconds, from the comfort of their own home.

What it does

We built a platform that leverages a microservices architecture to have individuals communicate with their desired physican model. The UI asks the user to input a picture then this picture is sent to the model for it to predict the type of illness. Then this value is returned to back to the user.

How we built it

We trained our models on datasets we found on Kaggle. We built our models using Python. We built the app using android studio. We sent requests via HTTP.

Challenges we ran into

We ran into challenges with getting the HTTP requests to run to and from the app and models.

Accomplishments that we're proud of

That we got the AI models to get very high accuracy even though we have a relatively small dataset.

What we learned

Connecting AI models to apps.

What's next for PhysicAIn

Adding more diseases detection models to the app. Also, if provided more time, we wanted to implement a chatbot that uses NLP to identify which illness is being studied.

Built With

• android
• java
• pil
• python
• scikit-learn