Stroke Guard

Inspiration

Stroke is a fatal medical emergency and has affected many people including people close to me over the past year. In the event of strokes, timely detection and immediate medical attention is critical in saving the person's life and any delay in detection can lead to permanent damage and could also lead to a fatality. The goal of stroke detection through video recognition of a person can help save a life in such events.

What it does

The application uses live video camera feed and does a prediction of the person in view of the video camera in event of a stroke and sends an email notification to persons/medical emergency response teams for immediate action.

How we built it

The web application was built using the Django framework and the prediction model uses a deep learning model that was trained using publicly available datasets. The prediction uses the trained model data to make predictions on the new live image data to give notification of the stroke prediction.

Challenges we ran into

The challenges we ran into was the integration of the frontend video data with the backend for the online prediction. The implementation of the deep learning model and appropriate data preprocessing was also a challenge.

Accomplishments that we're proud of

We were able to develop a web application that captures images online, and the offline prediction model that is able to predict the classification of stroke image