Using Inception Network v3 to Grade Diabetic Retinopathy Images
Date of Award
Spring 2020
Degree Name
Bachelor of Science
Department
Computer Science & Mathematics; College of Arts & Sciences
First Advisor
Vitaly Ford
Abstract
Diabetic retinopathy is a disease that causes abnormalities in the blood vessels of the eye that can lead to blindness. Doctors often use a grading scale from 0-4 to describe the severity of the retinopathy, but grading can be somewhat inconsistent between doctors and in some areas of the world it can be difficult to get regular screenings. This project explores using Convolutional Neural Networks, or CNNs, to automate the process of grading retinal images in order to make care for diabetic patients faster, more consistent, and more accessible. CNNs are deep learning algorithms that are often used in imaging because they excel at identifying important features in images. The goal of this project is to train a type of CNN called an Inception Network to categorize retinal images using the 0-4 grading scale.
Recommended Citation
Wallace, Anathae and Saeed, Muhammad Moiz, "Using Inception Network v3 to Grade Diabetic Retinopathy Images" (2020). Capstone Showcase. 2.
https://scholarworks.arcadia.edu/showcase/2020/comp_sci_math/2
Using Inception Network v3 to Grade Diabetic Retinopathy Images
Diabetic retinopathy is a disease that causes abnormalities in the blood vessels of the eye that can lead to blindness. Doctors often use a grading scale from 0-4 to describe the severity of the retinopathy, but grading can be somewhat inconsistent between doctors and in some areas of the world it can be difficult to get regular screenings. This project explores using Convolutional Neural Networks, or CNNs, to automate the process of grading retinal images in order to make care for diabetic patients faster, more consistent, and more accessible. CNNs are deep learning algorithms that are often used in imaging because they excel at identifying important features in images. The goal of this project is to train a type of CNN called an Inception Network to categorize retinal images using the 0-4 grading scale.