We wanted to try and create something both fun, useful, and a little silly. In particular, we wanted to create something that could spark social interaction by helping to break the ice, so we created an app that users can seek out other users as well as random people and use augmented reality to interact with them in the app, using dogs.

What it does

It's quite simple - your phone scans someone and determines if they're another player or not. If they're not, you're free to dunk on them as much as you like. If they are, they'll receive notifications when you dunk on them, and you'll have to have a dog of a similar or higher level to have a chance to dunk on them, otherwise they'll easily be able to dunk on you instead.

How we built it

We used mainly Python to build all the various components, using the OpenCV library for the face tracking and pygame for the game events.

Challenges we ran into

Using OpenCV in our project was very difficult, as it's a cumbersome API that requires a considerable amount of time to wrap your head around. Furthermore, having it done smoothly was an issue too. Ideally we'd have it as similar speeds to an iPhone face detection, but we were only able to optimize to a couple seconds rather than instant. We also ran out of time to properly deploy the app to a smartphone, but thankfully demonstrating on a laptop gets the point across just fine.

Accomplishments that we're proud of

What we learned

What's next for Dogdunk

Built With

Share this project: