Artificial intelligence can be put to good use in our society, and helping vulnerable groups get help when they need it most, in a automated efficient manner is a great step towards a smart-city and smart-society.
What it does
With easy integration of any internet-connected camera, Guardian Angel will constantly receive images of the video feed from said camera. Afterwards, the back-end will send this data to our Watson custom classifier, and receive a level certainty on whether or not the person has fallen. Our algorithm will then assess if someone is truly in distress. If this person does not get up, a Twilio text is sent to emergency services and contacts. Also, a user dashboard shows all your connected cameras for you to monitor.
How we built it
We built this project using the Watson API, Node.js for the back-end and Twilio.
Challenges we ran into
Our canvas (screen capture of webcam) was cropping our image when being sent to the back-end.
Accomplishments that we're proud of
Quickly trained an efficient A.I. with beautiful datasets (24 hours).
What we learned
We learned to use socket.io in relation with a database.
What's next for Guardian Angel
Face recognition for more personalized alerts to pre-set emergency contacts when in distress. User dashboard with authentication.