EyeQTest

Example Eye Test

EyeQTest

Submission for General Track

Many struggle with precisely measuring their eye score with convenience. To get a simple vision score, an individual would have to take time from their day to set up an appointment with their nearest Ophthalmologist and hope that the timing works out. Not only is the timing inconvenient for many people, the price can be too. In 2020, it was reported that about 50% of the U.S population has vision insurance, leaving over 100 million people to pay a premium. Further, even if an individual has vision insurance, the average price of a vision exam is approximately $95. For many Americans $95 dollars can be detrimental, but for the hundred million plus people without any vision insurance, it can cost them upwards of $250. Such extreme prices can lead many Americans to sacrifice such an important test for other things in life. This project tries to address and provide a solution for millions of people not just in America, but all over the world. EyeQTest seeks to provide a more affordable and convenient alternative to a typical vision exam that can still provide accurate results using only a laptop camera.

The hardest step in our vision test is the first step, being printing a template provided then most of the work is done internally. The template contains a simple box which will be used to see how far away the user is standing in reference to the camera. After the distance between the camera and the user is calculated, the letters will be displayed in a manner where it would be equivalent to seeing the letter twenty feet away. When the computer deems sufficient it will give the results, providing an accurate test for the user. Externally, the user should only do minimal work providing the most efficient and smooth process.

To provide a flowing experience, the computer must do all the hard work. When the user shows the template to the camera, the program will read the pixel length from corner to corner. Then it will compare the amount of pixels with a reference pixel count; using a ratio equation the program can commute how far the user is standing relative to the camera. For if Sally stands a distance x where the pixel length from each corner is measured to sixty pixels. The program will take the information and compare it to a reference image that is forty pixel length which would equate to ten feet. Subsequently, the program should do the math and compute the value of x.

If the project was fully realized, the computer would then list out a row of letters which the user would need to repeat out loud. In conjunction, the program will store the recorded data, and compare it to the right answers. If no errors are recorded the next row is displaced and so and so forth. The program will continue this process until the amount of wrong answers reach a certain threshold, which then the test will conclude. The results will then be compared to a vision test database, which then would give the accurate vision score.

The toughest part of the project was learning how to code for image recognition and django for the web server. The hurdle from knowing nothing to learning how to code for image recognition and django were the biggest leap. Thus, the project can be fully realized with the next hurdle being learning how to integrate audio input to make the experience flow better. This project takes all of us out of our comfort zone and helps build on our foundations.