Booth Id:
TMED033
Category:
Translational Medical Science
Year:
2021
Finalist Names:
Shah, Shaivi (School: Tesoro High School)
Abstract:
Problem Statement:
Cataract is the leading cause of blindness according to the latest assessment by the World
Health Organization (WHO). Majority of cases are from third-world countries where access to
professional tools and doctors is limited. However, traditional equipment have limitations such
as cost, regular maintenance, difficulty to use/interpret results without specialists.
Design Goal:
In order to try and solve these challenges, my objective is to design a low-cost
smartphone-based screening tool that identifies the risk for cataracts.
Procedures:
Oculi is made of two components: (1) a low-cost lens that allows anyone to take pupil
images (2) a smartphone application that displays a report of detailed cataract screening results.
In order to take the pupil images, the subjects were first taken in a dark room for the pupils to
dilate naturally. Then, the lens is placed 30cm/1ft from the subject’s eye to capture red reflex
images. Then, these images are fed to the application that uses machine learning to analyze the
images and find anomalies in the pupil.
Results:
After evaluating data, images obtained using the Oculi lens were 66% closer to gold
standard values. Images obtained using ophthalmoscope were 33% closer to gold standard
values. For the accuracy of the Oculi algorithm, after running 25 iterations on 256 images, the
model accuracy was 94%.
Conclusion:
Oculi is a two-part device that is portable, inexpensive, and simple to use. These
advantages can prevent the continual loss of vision notifying individuals the health of their eye
from early on.