For nearly two years, COVID-19 has negatively impacted the everyday lives of individuals around the world. While this may be due to many factors outside of our control, masks are an element within our control. If we could encourage the use of masks indoors through this alerting system, we can shorten the negative impact of this pandemic.
What it does
With the use of machine learning, our code trains itself through thousands of unmasked, face-up photos to classify when another image is masked or not.
How we built it
Using an open source computer vision library, we implemented its built-in functions alongside Python3 and a webcam to complete the task.
Challenges we ran into
As beginners to Hackathons, we had issues coming up with an idea that was feasible given the time frame. Even when we had the idea, we found trouble in sourcing sufficient training data to use for our facial recognition algorithm.
Accomplishments that we're proud of
Having only completed one previous Hackathon, our team has had very limited experience with coding on a strict time constraint. With that being said, we decided to jump off the deep end and tackle topics outside of our skill level.
What we learned
Through this project, we were able to get a foundation in an open source computer vision library along with an idea of how machine learning classification works.
What's next for MaskLink
In the short term, we are looking for a more consolidated way of notifying others about unmasked individuals. In the future, we hope to expand this technology to stores and airports where it would not just monitor mask wearing, but also be able to alert others of unwanted visitors.