Targeted Self Supervision for Classification on a Small COVID-19 CT Scan Dataset

Abstract

<p>Traditionally, convolutional neural networks need large amounts of data labelled by humans to train. Self supervision has been proposed as a method of dealing with small amounts of labelled data. The aim of this study is to determine whether self supervision can increase classification performance on a small COVID-19 CT scan dataset. This study also aims to determine whether the proposed self supervision strategy, targeted self supervision, is a viable option for a COVID-19 imaging dataset. A total of 10 experiments are run comparing the classification performance of the proposed method of self supervision with different amounts of data. The experiments run with the proposed self supervision strategy perform significantly better than their non-self supervised counterparts. We get almost 8% increase in accuracy with full self supervision when compared to no self supervision. The results suggest that self supervision can improve classification performance on a small COVID-19 CT scan dataset. Code for targeted self supervision can be found at this link: <a href="https://github.com/Mewtwo/Targeted-Self-Supervision/tree/main/COVID-CT" target="_blank">this https URL</a> </p>

Online Unsupervised Learning for Domain Shift in Covid-19 CT Scan Datasets

“…Our training strategy was a two-step process. For each model, we first performed targeted self supervision in a similar manner to Ewen and Khan [16]. We made horizontally flipped copies of our training images, and trained the network to determine whether an image was flipped or not.…”