Inspiration

Ive met many individuals with parkinsons and early detection can really help with symptom control

What it does

Uses a machine learning model to predict if an individual has parkinsons based off of vocal biomarkers

How we built it

I used a lstm+gru architechture and used the Mpower dataset where we turned audio intor spectrograms to be processed

Challenges we ran into

data was a huge challenge but after I got access to it computation became a large problem too but i used a paid google colab account which helped alleviate the computational strain.

Accomplishments that we're proud of

It WORKS!!!

What we learned

How to play with audio data and transform it into something useful

What's next for VoicePD

if I can have more compute power I would expand it and make it more accurate

Share this project:

Updates