Teyeland

The Solution
Optometrist Opinion
How It Works
Growth Drivers
Team

Inspiration

Our team members are myopic and hence we personally would like to find a better way to assess our myopia. Not to mention that it is a problem that affects 65% of grade 6 children, and 83% of Singaporean youths. (NA stats - nearly 74% of people in Canada need eyeglasses, contacts or laser eye surgery to see properly. About 41.6 percent of Americans are nearsighted) Additionally, our team aims to bridge the gap faced by traditional businesses in obtaining regular customers, as well as online eye-wear e-commerce websites that lack a simple and accurate way of obtaining a prescription online needed to secure a sale.

What it does

Our current MVP of Teyeland is an acuity check web app for eyecare providers powered by ML (Computer Vision and Voice Recognition). The user is able to follow the instructions on the screen to conduct a check on their vision and be recommended the severity level of needing to visit an optometrist based on their score.

How we built it

We first conducted a high level system design and feature planning for the overall application. We then did very basic mockups of the pages/components involved and explored how we intended the UX to be before further splitting the team into four units, specifically: UI, computer vision, speech to text, and scoring logic. We then synced regularly and troubleshooted issues each other were facing. Tech Stack: Next.js, Chakra UI, mediapipe, Deepgram

Challenges we ran into

We ran into issues getting mediapipe to work on next.js as well as integrating traditional react components to overlay the web camera used to conduct the acuity check.
Managing state flows was also a challenge as we had to sync the voice recognition model output with successful continuation of the acuity check.
We also had to ensure accurate distance estimation and head-angle tilt calibration to display the appropriate text sizing for the acuity check.
It was also worth noting that we had no prior background in optometry and our prior experience lies with our personal experience with myopia diagnosis, hence it was difficult to research into the pain points faced by both consumers and businesses in the eyewear & eyecare space and be able to come up with this solution for it.

Accomplishments that we're proud of

Amos: Setting up a finite state machine ⚙️
David: Picking up Chakra UI & building FE components 🎨
Xavier: Implementing real world research formula into logic scoring 🧮
Guo Jun: Implementing speech-to-text model 🎙️

What we learned

We have learnt to utilize cutting edge technologies such as Computer Vision and Voice Recognition in our solution, as well as learnt how to deploy such technologies onto a web app via mediapipe. We have also explored the use of React, Next.js, and Chakra UI to build the frontend application.

What's next for Teyeland

We will further develop the myopia diagnosis to include tests required to supply a prescription score as well as increase the accuracy of the diagnosis through the use of a ML model. We will also provide code snippets for quick implementation into online e-commerce eyewear businesses to close the gap in lost sales due to lacking prescription. We will also build a CRM Platform to provide marketing automation as well as forecast sales and predict churn of customers.