Inspiration
The chaos in the world right now. With the growing threat that is Corona Virus, we decided to develop a solution that can prevent the spread.
What it does
We use wristbands with Barometer, Gyroscope and Light sensor to track the movement of your arms and hands, then use advanced Machine Learning to detect whether the movement will result in the user touching their face and send a gentle reminder to them to wash their hands before.
How I built it
We scraped Android phone sensor data from the Chrome extension and then built a dataset of positive and negative samples of potential face touching movements. We trained a CNN to find these patterns.
Challenges I ran into
Fetching Sensor data, discretization of data, model training and implementation of it all.
Accomplishments that I'm proud of
The DL model, building the right training data, getting the data off the android phone
What I learned
Front end deployment of DL models using tensorflowjs and capturing sensor data using Chrome
What's next for Face Mat touch Corona
We plan to build custom lightweight hardware to accompany the software and distribute it across the world as soon as possible.
Log in or sign up for Devpost to join the conversation.