IRYS

With ADHD in children rising every year, there needs to be a solution to combat this. Introducing: IRYS, an ADHD diagnosis program using eye-tracking and machine learning for more accurate prognosis.

Comment

Inspiration

  • One of the members in our group, Rahul Ramakrishnan, has had ADHD for his entire life, yet was misdiagnosed multiple times due to his other health conditions. For example, his sleep apnea was said to be the culprit of his lack of concentration, yet this was not the case. Being misdiagnosed for years caused him to fall behind academically and socially. So many other children experience what Rahul experienced everyday (until he got treated properly), hence we felt this project has the potential to change millions of lives. And with little methods to diagnose one of the most common learning disabilities in the world, we took the opportunity and seized it to create a one-of-a-kind eye tracking software.
  • We wanted to help solve a prominent issue Health Care issue and with ADHD still being a relevant issue, we decided to try and use technology to solve this problem. We also wanted to make use of Artificial Intelligence in some fashion to ensure the best results.

What it does

  • With ADHD rates increasing in children along with a lack of proper diagnosis, something must be done about this global issue: Introducing IRYS, an ADHD diagnosis tool designed to track how distracted a user is by analyzing eye movements. The application tracks the user's eyes over a period of time and calculates whether the user is concentrated on the screen. The system then reports the data over to CockroachDB to be used to identify overall trends. IRYS stores data on a database, and outputs a user-friendly visualization for further analysis.

How it works

  • Start IRYS @ https://irys.vercel.app/
  • IRYS runs in the background, using your webcam to analyze eye movements.
  • IRYS stores data on a database, and outputs a user-friendly visualization.

How we built it

We built the webapp using NextJS as the frontend and a variety of other tooling and APIs for the backend. The majority of the project is built using Node.js which is the main way of connecting the front and back end and Python to communicate with the database and run the machine learning algorithms. The incentive for picking NextJS is the power of server side rendering because we had primarily data that was driven by the backend, therefore we made use of a framework that best supported this philosophy. We also wanted to do web development because it is the most accessible and fastest way to reach a lot of people at once.

Challenges we ran into

  • Interacting with the database
  • Communicating between devices
  • Eye tracking software
  • Lighting
  • Tracking eye gestures instead of movements
  • Started off on Nvidia Jetson Nano, moved over due to network issues
  • Page Styling

Accomplishments that we're proud of

  • Application ran smoothly with no errors
  • Deployed correctly
  • Integrated multiple programming languages
  • Able to call a database
  • Used hardware and software together
  • Utilized machine learning algorithms
  • Built our own API
  • Clear and usable UI
  • Connected technology with healthcare

What we learned

Throughout this project, we learned a magnitude of things. Firstly, ADHD is one of the most complicated disorders, going beyond simply “a lack of concentration”, and has a ripple effect on its damages in all aspects of life. We also learned that there are many ways to approach this problem, and an eye tracker is not a common solution. The most technological solution comes from EEG’s, which measure alpha and beta waves in the brain when determining focus and attention. Other things we learned: time management, splitting the workload, how to approach difficult problems, connection between your front and back integration, using machine learning algorithms, how to make a visualization in python, and a multitude of other skills.

What's next for IRYS

While we feel this is a great product, we believe with extra research and development, funding, and time, this can become the standard for future ADHD testing with pinpoint accuracy. First of all, however, is accessibility. We hope to be able to implement this product into smartphones, as cameras are improving in quality every year and some even surpass that of sophisticated web cams. This way, home tests become more of a reality, as most of the population has access to a basic smartphone. Additionally, adding more eye patterns into the database of data can make our program process and document more accurate eye movement that indicate a state of zoned off or distraction. We also wish to implement a way to track zoning off, like adding a counter to our distraction variable every time the participant stares at a specific location for longer than intended. Also, other learning disabilities like dyslexia have some connection to eye movements, hence we can tweak the program to cover more dyslexia behavior to cover a wider range of learning disabilities. Aside from this field, we can also expand the eye tracking software into the education sector, creating monitored tests that track eye movement for academic honesty. In reality, the sky’s the limit in terms of the potential for our product, but these are the main ideas we thought of to expand this product.

Built With

Share this project:

Updates