Deep Learning for Computer-Aided Diagnosis in Ophthalmology: A Review

James Brown*, Georgios Leontidis

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Artificial intelligence has once again come to the fore following the tremendous success of one class of mathematical models, the artificial neural network. The newly coined approach “deep learning” has dominated modern scientific discourse, infiltrating the fields of physics, chemistry, engineering, biology, and medicine. One of the most talked-about applications of deep learning is in computer-aided diagnosis (CADx), having a profound influence in the medical branches of radiology and ophthalmology. This chapter presents and discusses some of the recent developments in CADx with applications to ophthalmology, beginning with the motivation and historical approaches to retinal image analysis, providing insight into the limitations of previous methods. This leads to a discussion about modern solutions and state-of-the-art, highlighting various caveats that may curtail clinical adoption. The chapter goes on to discuss recent developments in the field of machine learning, offering speculation about future directions for CADx in ophthalmology, and how one might address the most debated critiques of contemporary deep learning solutions.
Original languageEnglish
Title of host publicationState of the Art in Neural Networks and Their Applications
EditorsS. El-Baz, Jasjit S. Suri
Place of PublicationLondon, UK
PublisherElsevier
Chapter11
Pages219-237
Number of pages19
Volume1
Edition1st
ISBN (Electronic)9780128218495
ISBN (Print)9780128197400
DOIs
Publication statusPublished - 21 Jul 2021

Keywords

  • Computer-aided diagnosis
  • opthamnology
  • visual impairment
  • deep learning
  • medical image computing
  • optic neuritis
  • glaucoma
  • diabetic retinopathy

Fingerprint

Dive into the research topics of 'Deep Learning for Computer-Aided Diagnosis in Ophthalmology: A Review'. Together they form a unique fingerprint.

Cite this