b# | Devpost

Home Screen
Sleep quality averaged over one week

Inspiration

Hackathons, of course!

What it does

B# uses IBM Watson's facial recognition and machine learning to detect your level of sleep deprivation. Only 11% of students get a healthy amount of sleep; in our day and age, our work has led us to an unhealthy and destructive sleep cycle that is not conducive of learning. It also keeps track of your sleep cycle within the last 7 days, providing an intuitive representation of your sleep deprivation levels to promote a more healthy lifestyle. Sleep deprivation is the root cause of many easily avoid able accidents - 30% of accidents occur because a driver falls asleep at the wheel.

How I built it

We integrated machine learning and image recognition API from IB Watson algorithms on our Ionic-based app and Node.js based website to predict the trends. We "trained" our machine by pairing the "scores" of images of sick people with varying degrees of drowsiness (including members of TreeHacks), allowing it to diagnose sleep deprivation based on the "scores".

Challenges I ran into

For most of us, this was the first time we tackled machine learning. Implementing the IBM Watson API in our hack was a particular challenge, especially getting it to speak to our app and website.

Accomplishments that I'm proud of

Based on user feedback, B# is successful in estimating the tiredness 92% of the time and provides an its intuitive interface provides for an easy platform to improve one's overall health.