A post on DevRant by a hearing impaired developer who describes the impact of technology and automation in improving quality of life for individuals with disabilities
What it does
A wristband that detects sirens in real time, and alerts the wearer of their presence. The target audience is hearing impaired individuals who drive.
How we built it
We use the Urban Sound Dataset to train a neural network to recognize the sound of sirens. The application then signals the Myo wristband which vibrates to alert the wearer. We used Python bindings for the Myo SDK.
Challenges we ran into
Improving model performance, working with corrupt data, microphone issues, training in AWS, attempting to port application to Android
Accomplishments that we're proud of
Getting a prototype working; understanding how neural networks work as beginners to machine learning;
What we learned
Interfacing with software. AWS
What's next for Siren Detector
Move all functionality to a smart watch; extend to detect other life threatening sounds like gunshots, trucks backing up, fire alarms, etc.