BlinkAI
🏆 Best Beginner Hack
February 2025
PythonFlaskJavaScriptOpenCV
About This Project
BlinkAI came out of a hackathon idea: what if someone who couldn't use their hands could still communicate just by blinking? I built a system that watches your eyes through a webcam, picks up on blink patterns, and turns them into Morse code, then into text — all in real time, with about 95% accuracy.
I used OpenCV to track the face and detect blinks, then wrote the logic to turn different blink lengths into dots and dashes. The whole thing runs through a simple web interface I built with Flask and JavaScript so anyone could try it without installing anything.
It ended up winning Best Beginner Hack at the hackathon, and The Stanford Daily wrote an article about it, which was pretty surreal since it started as a project I wasn't even sure would work.
Technical Challenges
Getting blink detection to actually work in real time without lagging, dealing with different lighting messing up the camera tracking, and making the interface simple enough that someone could use it without instructions.
What I Learned
This was my first real project with computer vision, so I learned a ton about OpenCV and how finicky face tracking can be. Also got a lot faster at building stuff with Flask.
Project Goal
Built an assistive communication tool that translates eye blinks into text using computer vision and Morse code.
Engineering Focus
Focused on real-time facial tracking, blink detection accuracy, Morse code translation, and accessible interface design.
Skills Used
Python, Flask, JavaScript, OpenCV