Think back to your wedding or someone else’s wedding you’ve been to the past. There’s always one guy who can’t dance but just doesn’t care. Usually on old tv shows it’s someone’s dad and with each second he’s losing reputation with friends, enemies watch smugly from a far, no one wants to be next to him. It’s because he’s making himself unattractive.
What it does
We brought the essence of dad dancing to you. First we’re targeting people who seriously want to be away from others. Introverts. Next using machine learning to do sound classification we’ve written an app to identify human voices. When the app detects a voice it starts vibrating and optionally playing a sound or song. We do this for 3 reasons:
- It disrupts someone from being around you and alerts you to them.
- We can aggregate this for statistical purposes and thus gamifying social distancing (like your car telling your your average mpg)
- For shock value. We want this to be memorable to help the app spread like a meme.
How we built it
We used python to do our analysis and create our models along with Apple's ML API. We wrote the iOS app in Swift.
Challenges we ran into
- Learning Swift.
- Gathering enough audio samples for training a model.
- Training audio models.
Accomplishments that I'm proud of
We've never made an iOS app before so it was a great challenge to overcome in the short amount of time. We had to transfer our preprocessing/normalization methods which worked on text to audio files and on top of it write it sometimes in swift.
What we learned
- Audio preprocessing
- Audio Machine Learning methods ## What's next for Vamoose