BreatheSafe

Inspiration

October is SIDS Awareness Month, and our team wanted to build something meaningful that could help prevent one of the most heartbreaking and overlooked issues in infant care - unsafe sleep and breathing irregularities.

Our inspiration came from real experience. Team members Kalyan, an EMT, and Kris, who works in healthcare, have seen the effects of unsafe sleep environments firsthand. As a team, we’re all passionate about health, safety, and technology, and we wanted to combine those worlds to make a positive impact.

We’re a diverse group of USC students and alumni, from undergraduate to graduate to post-grad, who didn’t know each other before the hackathon. Within 24 hours, we bonded, aligned on a shared mission, and built something we’re proud of: BreatheSafe.

What it does

BreatheSafe is an AI-powered platform that can be integrated with baby monitors that uses computer vision and machine learning to track breathing, detect unsafe sleep positions, and deliver real-time safety insights to parents through a mobile app.

The app also includes AI-driven recommendations, historical breathing trends, and educational resources about safe sleep practices - all designed to help parents monitor and understand their baby’s well-being with peace of mind.

How we built it

We built a Python backend using MediaPipe, OpenCV, and scikit-learn to process video frames, analyze breathing patterns, and classify safe vs. unsafe sleep positions in real time.

A React Native frontend (built with Expo) connects via WebSocket, visualizing live data, breathing rate, and position confidence levels. The app also integrates OpenAI’s GPT-4 API to generate personalized safety advice and insights.

We collaborated through GitHub, built and tested components locally, and integrated everything in under 24 hours.

Challenges we ran into

The biggest challenge early on was agreeing on a direction. We came from different fields: tech, health, and business. Once we settled on the problem of infant sleep safety, our collaboration clicked.

We also faced technical hurdles in synchronizing real-time video analysis with the mobile app feed, managing latency, and refining the breathing signal filtering to reduce false readings.

Accomplishments that we're proud of

We’re proud that within 24 hours, we built a fully functional prototype that combines AI, computer vision, and healthcare impact, all while working together for the first time.

We’re especially proud of how quickly we bonded as a team, leveraging each of our strengths to create something that could truly help parents feel more secure about their baby’s safety.

What we learned

We learned how to integrate AI/ML, real-time computer vision, and mobile UX into one cohesive system. We also learned how powerful collaboration can be when people with different expertise, from healthcare to AI to design, unite around a shared goal.

This project reminded us that technology can be both intelligent and compassionate - and that even in a short hackathon, innovation can come from empathy.

What's next for BreatheSafe

We plan to expand BreatheSafe by:

• Deploying the backend to the cloud for remote monitoring
  
• Supporting multi-baby environments like daycares
  
• Adding predictive analytics for early warning patterns
  
• Integrating with smart nursery devices
  
• Conducting clinical validation to move toward real-world safety use
  

BreatheSafe isn’t just a hackathon project - it’s our vision for safer, smarter baby sleep monitoring powered by AI.

Built With

• chart-kit
• expo.io
• joblib
• mediapipe
• numpy
• openai-api
• opencv
• python
• react
• react-native
• react-native-svg
• react-navigation
• scikit-learn
• scipy
• streamlit
• websocket
• websockets