Sweet Beats

Inspiration

We were inspired by the loud music in the auditorium after the initial presentations.

What it does

It uses pose tracking on multiple people in browser to check when there is dancing in the room. If people are dancing the music plays. The harder you dance the louder it plays and the more emphasis is placed on the visuals.

How we built it

It uses a tensor flow javascript pose tracker called potent to run in browser. Since we were unable to get a Kinect for the project, it uses an estimation based on the webcam without taking depth into account. The more motion people create, while being in view and above a certain "dancing" threshold, the louder the music plays and the greater the visuals.

Challenges we ran into

It was hard to get both the pose tracking and the visuals to work nicely. We debugged both for hours. We attempted to run a playlist through Spotify, but were unable to decipher the Spotify API in enough time.

Accomplishments that we're proud of

Pose tracking is really cool! Especially running in real time in browser! Additionally we are proud of getting those sick visuals to work.

What I learned

We learned a lot about how to use a webcam through the browser and doing pose tracking without a depth camera.

What's next for Sweet Beats

We want to add Spotify integration! It's definitely possible, just needs a bit more time.

Built With

• html5
• javascript
• tensorflow