ConcussionSite

Instant Interactive Concussion Screening—on the go, on the site, save your life.

Comment

Inspiration

A close friend of ours took a hit during a basketball game and walked it off, claiming he was "probably fine". We kept telling him to get checked up but like most athletes, he did not want to deal with the issue. Over the next few days the symptoms got much worse and he finally went to the doctor, which diagnosed him with a concussion.

This experience made us realize how common it is for our friends that play sports to have no proper way to ensure how serious a head injury is. As of now there is no quick proper way to screen a concussion, and we wanted to change that.

What it does

Concussion site is an app based concussion screening tool that runs real time neurological tests using only your webcam.

It asses key concussion symptoms by guiding users through:

  • blink response, flickering under light
  • Smooth purist eye tracking with a moving dot
  • Gaze stability and fixation tests
  • Blink rate and irregular eye movement detection
  • short symptom questionnaire

All these results are instantly analyzed and summarizers to an easy understandable report. Users can also ask questions about their results as well as be able to send an email through the app to the doctor with the provided results.

How we built it

We built ConcussionSIte using: MediaPipe + OpenCV for face mesh, eye tracking, blink detection, and real-time landmark analysis Custom Python logic for blink frequency, photophobia responses, and pursuit accuracy A React/Google App Script-style frontend (or your actual stack) for guiding user tests in a clean UI Gmail OAuth for secure automatic report emailing Cursor AI + iterative agents to build the UI flow, integrate tests, and support interactive report editing

Challenges we ran into

Getting blink detection and the other tests to work with eye scanning. Integrating an agentic system with the response allowing for user interaction Integrating Gmail OAuth cleanly into the project Aligning UI and the testing in a smooth workflow Creating tests that feel medically meaningful without claiming diagnostic authority

Accomplishments that we're proud of

Building a fully working concussion-screening prototype in a short time Successfully integrating real-time eye tracking through a webcam Designing a clean, simple UI that non-technical users can follow Making a tool that could genuinely help people/athletes know when to seek medical attention Implementing automated report generation + optional email delivery

What we learned

How to combine modern computer vision tools with real clinical screening concepts How to asses normal human responses for potential damages. How important it is to design health-related tools with clarity, transparency, and safety How to collaborate quickly using AI tools (Cursor, GitHub, etc.) to ship features fast How small sensor-based features can create big impact in real-world safety

What's next for ConcussionSite

Adding saccade tests, memory recall tasks, and reaction-time tests Further Developing/Deploying a mobile version so athletes can use it on the sidelines Making it more personalized for the user (accounts, etc.) Creating a full athlete baseline profile so results can be compared over time Partnering with schools, rec leagues, and trainers for real-world testing Integrating with Apple Health / Google Fit

Built With

Share this project:

Updates