Imagine your favorite piece of music playing under a super annoying drowning, buzzing, This is what its like for many wonderful people afflicted with Tinnitus.
Its a personal issue for me that needs a modern and effective solution. My grandma developed age-related tinnitus where she has a perception of noise and sometimes ringing in the ears.
Our solution, that we are extremely excited about, is using machine learning to effectively balance out the perceived noise in a continuously improving fashion. Let's make the music ring, not Tinnitus!
What it does
Buzzless is a use a Deep Learning model to extract clean speech from noisy audio with high accuracy, thus achieving denoising
How I built it
We trained and evaluated the Deep Learning model on NSDTSEA and VOiCES (created by Lab41) datasets. The model is based on the architecture introduced by François et. al. in "Speech Denoising with Deep Feature Losses".
Challenges I ran into
Accomplishments that I'm proud of
Yusuf is a first time hacker.
What I learned
So much about audio processing and deep learning.
What's next for Buzzless
Taking our solution to the masses! The aging population of Canada is an amazing market for our tech. The goal is to produce a prototype for trial tests within 4 months. Our Vancouver based accelerator is able to launch new tech solutions from inspiration to implementation within that time. Let's make the music happen!
Technical video demo: https://youtu.be/6cFNEXny6dI