/OCT-Scans-Classification

This project compares two learning paradigm, namely transfer-learning and self-supervised learning in a classification task of three retina disorders CNV, DME and DUSEN in addition to the normal condition using an OCT B-scans

Primary LanguageJupyter Notebook

Project title:

Retina Disorders Classification via OCT Scan: A Comparative Study between Self-supervised Learning and Transfer Learning

project group members:

Name ID Email
Saeed A. Shurrab 135325 sashurrab18@cit.just.edu.jo
Yazan W. Shannak 137027 ywshannak19@cit.just.edu.jo
Rehab Duwairi Supervisor rehab@just.edu.jo

Abstract

Retina disorder is among the common types of eye disease that occur due to several reasons such as aging, diabetes and premature born. Besides, Optical Coherence Tomography (OCT) is a medical imaging method that serves as a vehicle for capturing volumetric scans of the human eye retina for diagnoses purposes. This research compared two pretraining approaches including self-supervised learning (SSL) and transfer learning (TL) to train ResNet34 neural architecture aiming at building computer aided diagnoses tool for retina disorders recognition. In addition, the research methodology employs convolutional auto-encoder model as a generative SSL pretraining method. The research efforts are implemented on a dataset that contains 109,309 retina OCT images with three medical conditions including CNV, DME, DRUSEN as well as NORMAL condition. The research outcomes showed better performance in terms of overall accuracy, sensitivity and specificity 95.2%, 95.2% and 98.4% respectively for SSL ResNet34 Model in comparison to scores of 90.7%, 90.7% and 96.9% respectively for TL ResNet34 model. In addition, SSL pretraining approach showed significant reduction in the number of epochs required for training in comparison to both TL pretraining as well as the the previous research performed on the same dataset with comparable performance.