DrinkWise

Inspiration

The inspiration for DrinkWise came from the devastating effects of drunk driving and the need to address this problem on college campuses.

What it does

DrinkWise is a solution that combines both facial recognition and audio recognition technology to determine an individual's level of sobriety. This allows us to check if they are fit to drive a car.

How we built it

DrinkWise uses a custom-built neural network for its facial recognition system, which is trained to detect signs of inebriation, such as dilated pupils, unsteadiness, and an overall disheveled appearance. Its audio recognition technology transcribes spoken words and compares it to an original phrase to identify any disparities in pronunciation, intonation, or enunciation, which can serve as an indicator of possible inebriation.

Challenges we ran into

We ran into a TON of bugs, the main one being data. Since we built the networks from scratch we had to find our own data, and there wasn't much available online, but we were able to find a few obscure data points and compile a training set.

We also had a hard time linking our networks to our website, but through countless restless hours, we figured it out!

Accomplishments that we're proud of

DrinkWise is designed to be mobile responsive, accessible, and convenient, making it easy for users to utilize the technology on the go, anywhere and anytime - and we achieved that.

What we learned

When creating the neural networks - we learned a TON about theory and application into the world of neural networks. We also learned about a ton of algorithms like the Levenshtein distance algorithm.

What's next for Drink Wise

The goal of DrinkWise is to make a difference in the world and save lives by preventing drunk driving incidents. The creators hope to reduce the number of drunk-driving accidents and save lives, not just on the UC Santa Cruz campus but across the globe. We plan to achieve this with a mobile app & user feedback.