Classification of Multiple Retinal Diseases using Vision Transformer

  • Manuel Alejandro Rodriguez Rivera

Student thesis: Master's Thesis

Abstract

Early detection of retinal diseases is one of the most important means of preventing partial or permanent blindness in patients. In this research, we propose a novel multilabel classification system for the detection of multiple retinal diseases, using fundus images collected from a variety of sources. First, a new multi-label retinal disease dataset is generated, by combining a number of publicly available datasets for fundus disease classification. Next, a sequence of post-processing steps is applied to ensure the quality of the image data and the range of diseases, present in the dataset. For the first time in fundus multi-label disease classification, a transformer-based model is proposed for image analysis and decision making. Numerous experiments are performed to optimize the configuration of the proposed system. We show that the approach performs better than state-of-the- art works on the same task by 8% and 7.7% in terms of AUC score for disease detection and disease classification, respectively. The obtained results further support the potential applications of transformer-based architectures in the medical imaging field.
Date of AwardMay 2022
Original languageAmerican English

Keywords

  • multi-label
  • fundus imaging
  • disease classification
  • transformer
  • deep learning.

Cite this

'