ConCussify - The Idea

Our submission for the Ideathon Division

Inspiration

Concussions are known for going undiagnosed, particularly in sports, where there’s pressure to play through injury and more obvious symptoms may not immediately be evident. My sister, for instance, experienced multiple concussions as a competitive cheerleader without realizing, and this resulted in her receiving delayed medical attention and suffering through worsened symptoms. We were motivated to address this problem by developing a convenient and accessible solution for athletes and other individuals prone to concussions, empowering them to independently assess their condition on the spot and seek medical help promptly. By combining a user-friendly and interactive interface with AI features to fill-in for potentially unavailable healthcare professionals, our web app aims to make concussion diagnosis more accessible and efficient than ever before.

What it does

Concussify determines the likelihood of a concussion based on whether it detects a difference in the pupil size of the right and left eye (a common indication of a concussion). We had also planned for users to be able to take a series of baseline tests that can be repeated after a head injury to determine if there’s been a change in their cognitive speed, memory, or awareness.

How we built it

The frontend was written in CSS and HTML, and we developed the backend using Cloudflare Workers, which is an edge-based serverless provider that is based upon the WinterCG runtime; through this, we are able to create a highly-scalable and mobile backend able to tackle almost any challenge. We do not have any lock-in as it is based on the Winter CG Runtime which allows us to move to other providers such as Vercel or Deno. We decided to implement Google sign-in to reduce development time while still providing ease-of-use. The pupil size detection function was built using two pretrained machine learning models. The image is first processed through OpenCV’s Haar Cascade model for eye detection, then the separate eye images are

Challenges we ran into

We faced a number of challenges, from differing timezones to part of the team being constrained to work on it for a day and a half. The communication issues which the differing timezones did not help with, nevertheless we still tried our hardest to overcome this obstacle and are proud of the accomplishments we achieved.

Accomplishments that we're proud of

We are proud of the lesson’s we’ve learned, the challenges we tackled and our ability to still able to make something in a short timeframe. For some this was their first experience in working in a hackathon, through this they have began their journey in participating in many more hackathons like this one. From being able to work in a fast paced stressed environment to completing challenges that we’ve never faced before, being able to see it through to the end is what completes us and we are proud to be able to accomplish this through teamwork.

What we learned

We learned a lot through this hackathon, got to work with new technologies, and it was very fun. Our team got to experience new fields we weren’t familiar with, and we enjoyed it.

What's next for Concussify

We hope to extend on our AI Model, which due to time limits we weren’t able to fully finish and connect to our main app. We also think Concussify could have a lot of improvements and adjustments as well.

Built With

• video