Inspiration
Drowsy and distracted driving is a major cause of road accidents, often leading to severe consequences. We wanted to create a real-time AI-powered solution that helps drivers stay focused and safe by detecting when they look at their phone or close their eyes for too long.
What it does
Guardian Angel continuously monitors the driver’s face using computer vision to detect signs of distraction or drowsiness. If the driver looks at their phone or closes their eyes for too long, the system sends an alert to help them refocus on the road.
How we built it
We used Next.js/React for the frontend, providing a smooth and responsive UI. FastAPI powers the backend, running an OpenCV-based AI model that processes video feeds to track eye direction and detect when eyes are closed. We integrated WebRTC to efficiently stream video from the user’s device to our backend for real-time analysis.
Challenges we ran into
Optimizing real-time video processing to minimize latency while maintaining accuracy. We originally were sending images from the web camera in intervals, causing extreme latency. We moved to using WebRTC to stream the video directly.
Accomplishments that we're proud of
Successfully implemented real-time distraction detection with minimal delay. Built an intuitive and lightweight web app that runs efficiently on different devices. Integrated AI, WebRTC, and FastAPI into a cohesive system within a short timeframe.
What we learned
Best practices for optimizing WebRTC for real-time video streaming. The importance of balancing performance and resource efficiency in AI-driven applications.
What's next for Guardian Angel
We plan to enhance Guardian Angel by:
Improving model accuracy with more training data using fine-tuning. Developing a mobile app version for seamless use on the go. Exploring integrations with smart car systems for a more robust safety solution.
Log in or sign up for Devpost to join the conversation.