What it does
Eagle Eye is a smart AI camera that detects weapons and notifies all users of a potential threat in real-time.
How we built it
We created the client side with React and leveraged tensorflow.js to provide real-time weapon detection. For the models, we trained them on IBM Cloud using object storage and IBM Watson. After that, we created a backend server with socket.io to handle real-time data updates and updates all clients simultaneously when suspicious activity is detected.
Challenges we ran into
Deploying to GitHub pages and Heroku. Yarn difficulties. React components not working. Not enough time to host on AWS.
Accomplishments that we're proud of
Being able to train the model and build a full-stack application in a single day.
What we learned
Mark: Learned how to train AI models using IBM's Cloud Annotation UI.
What's next for Eagle Eye
Deploying it to production. Further training the models to improve accuracy and detect a variety of weapons.