Skip to content

ACM Research 2020 | A convolutional neural network to classify retinal optical coherence tomography scans.

Notifications You must be signed in to change notification settings

ACM-Research/image-classification-cnn

Repository files navigation

Convolutional Neural Network to Classify Retinal OCT Images (Fall 2020)

Poster

Image of Research Poster

Objective

Optical Coherence Tomography is a Retinal imaging technique used to detect retinal abnormalities. However, the analysis of such images to reach aretinal diagnosis serves to be a difficult and time-consuming task for medical professionals. This project aims to develop the most efficient convolutional neural network model to classify these OCT images. This CNN is then benchmarked against a fully-connected neural network model to demonstrate a CNN's performance in tasks that require spatial context.

Background

OCT is a retinal imaging technique which uses light waves and their reflections to build a cross-section of the eye. This cross-section image is used to look at the different layers of the retina and identify any signs of abnormalities to diagnose optical diseases. For the purposes of this project, we focused on 4 types of classifications – Choroidal Neovascularisation (CNV), Diabetic Macular Edema (DME), Drusen present in Age-related Macular Degeneration (DRUSEN), and a Normal Retina (NORMAL). Figure 1 depicts an OCT Image of each of these abnormalities, along with a normal retina for a comparison

OCT Image Example

Models

Our models are constructed through Keras, a Python library used for deep learning. We experimented with different sets of layers and measured our model's effectiveness using accuracy, F1-score, and sparse categorical cross-entropy loss metrics. The models are compiled with their respective training sets to produce the required metrics.

Model Data

Online Demo

Demo: https://colab.research.google.com/github/ACM-Research/image-classification-cnn/blob/master/demo/RetinalOCTDemo.ipynb

Images to use: NORMAL.jpeg, CNV.jpeg, DRUSEN.jpeg, DME.jpeg

Contributors

About

ACM Research 2020 | A convolutional neural network to classify retinal optical coherence tomography scans.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published