Helping computer users manage their eye health during the pandemic while staying productive.
What it does
Using the webcam's live video feed, it detects when the eyelids are open, half, closed to count number of blinks, speed of blinks to provide you with realtime data about your blinking habits that may be causing discomfort such as dry eyes or eye strain or longterm vision problems such as myopia.
How we built it
Frontend: MERN stack. Backend: Python (Flask) server + OpenCV (for facial detection) Non-tech: Blood, sweat, and tears of devs: Cat, Jessica, Jimmy, Rhys.
Challenges we ran into
Making a good UI was a challenge for us since we consist of 4 developers, but with good communication and ad hoc meetings we made it through. There were many bugs (as expected), but notably installing dlib on windows to use OpenCV was a very time consuming one.
Accomplishments that we're proud of
Accurately detecting a blink, keeping a counter of it, tabulating the blinks per minute, identifying poor eye health and providing interactive recommendations to alleviate symptoms!
What we learned
We learned more about the power of Computer Vision through OpenCV and how we can use it to improve a person's eye health. Additionally we learned more about creating a web app and making server requests between a MERN frontend and a Flask application. By researching articles and papers regarding eye health, this was great for our own personal benefit too!
What's next for iBlink
iBlink's current state is good foundation for detecting possible eye symptoms that relate to poor eye condition. The next steps with iBlink would be to detect more complex behaviours and symptoms such as squinting and redness in eyes. We plan to also give a more detail summary of the behaviour and symptoms to the user to improve the diagnostic on a person's eye health.