SimpliSteth

Inspiration

I initially wanted to make an app that measures heart rate through audio, but it wasn't audible through a phone microphone, but I realized while listening to the audio, that I could measure lung health through breathing.

What it does

It is supposed to measure your breathing for crackles and wheezing by placing the microphone onto your throat.

How we built it

I modified a pre existing tensor flow audio prediction model to detect crackling and wheezing, and then placed it into a pre existing android studio app, for the sake of time.

Challenges we ran into

The audio classifier I used was actually meant for transfer learning, and by the time I had realized this it was too late, so the app mostly predicts what that audio classifier had already been doing. Due to this the app remains unfinished.

Accomplishments that we're proud of

I liked my graphic designing and idea making.

What we learned

I learned that team members are crucial in competitions and so is background knowledge in the software you use.

What's next for SimpliSteth

I would have added a function that replays your audio recording back to you for manual checking as well. I would also replace the whole model to something that does not use transfer learning, and replace the app ui to something custom.

Built With

• java
• python