Inspiration

The average university student is constantly staring at their laptop or their iPad, focused on lecture slides, or a practice exam, or some form of content. These long hours of studying and screen usage commonly lead to eye strain or fatigue, however not many people are actually aware of how little they blink while concentrated on a screen. Ample rest is also often impacted by device over usage, so the program alerts users of their yawn count.

What it does

VisioScan uses access to the user's laptop webcam in order to track their blinking and display their blinking rate. If the blinking rate falls under the threshold of 10 blinks/minute, the program delivers an on-screen notification that this may be an indication of possible fatigue or high concentration. Bad screen habits are a universal student experience, and we wanted to build a simple tool that can recognize and share awareness.

How we built it

We built Visioscan through the usage of Python, OpenCV, and MediaPipe FaceMesh to obtain facial landmark tracking. The eye aspect ratio allowed us to compute eye openness, which was developed into the blink detection aspect. A yawning detecting was additionally implemented with the same conditions of a mouth ratio aspect.

Challenges we ran into

We faced difficulty integrating various hardware components into VisioScan.

Accomplishments that we're proud of

We are proud that we persevered and were able to quickly problem solve under stress and time constraint.

What we learned

The initial project idea was a bit too ambitious, and out of the scope of our skills. We learned how to quickly devise a new plan under limited by building off our initial goal of a healthcare or wellness oriented solution.

What's next for VisioScan

Moving forwards, we hope to develop VisioScan so it can read various emotions, detect signs such as lack of sleep, and symptoms of intoxication. We aim for this tool to be able to read the entire human face. In addition, we want this to be able to read more than one face at a time.

Built With

Share this project:

Updates