Diving Deep into Deep Learning: an Update on Artificial Intelligence in Retina
- PDF / 271,086 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 14 Downloads / 198 Views
RETINA (R GOLDHARDT, SECTION EDITOR)
Diving Deep into Deep Learning: an Update on Artificial Intelligence in Retina Brian E. Goldhagen 1,2 & Hasenin Al-khersan 1
# Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Purpose of Review In the present article, we will provide an understanding and review of artificial intelligence in the subspecialty of retina and its potential applications within the specialty. Recent Findings Given the significant use of diagnostic imaging within retina, this subspecialty is a fitting area for the incorporation of artificial intelligence. Researchers have aimed at creating models to assist in the diagnosis and management of retinal disease as well as in the prediction of disease course and treatment response. Most of this work thus far has focused on diabetic retinopathy, age-related macular degeneration, and retinopathy of prematurity, although other retinal diseases have started to be explored as well. Summary Artificial intelligence is well-suited to transform the practice of ophthalmology. A basic understanding of the technology is important for its effective implementation and growth. Keywords Artificial intelligence . Machine learning . Neural networks . Diabetic retinopathy . Age-related macular degeneration . Retinopathy of prematurity
Introduction to Artificial Intelligence Artificial intelligence (AI) has and will continue to transform the field of ophthalmology, particularly within the subspecialty of retina. Many techniques have been employed to achieve the goal of machines acting “intelligently” like humans [1]. While some methods utilize the programming of explicit instructions, the technique of machine learning may also be employed. In machine learning, the computer makes inferences through the identification of patterns or features from
This article is part of the Topical Collection on Retina * Brian E. Goldhagen [email protected] Hasenin Al-khersan [email protected] 1
Department of Ophthalmology, University of Miami Miller School of Medicine, Bascom Palmer Eye Institute, 900 NW 17th Street, Miami, FL 33136, USA
2
Miami Veterans Administration Medical Center, 1201 NW 16th St, Miami, FL 33125, USA
data. This learning can be performed using a system of layered algorithms that work in an interconnected fashion, called a neural network, which gets its name from its weak similarity to the organization of neurons in the human brain [2]. Each algorithm in the layer works on recognizing particular features and weighs them with regard to importance prior to feeding an output to the next algorithm layer [3]. When there are many such layers within a network, it is considered to be “deep learning.” While learning may be supervised, unsupervised, or semi-supervised, the vast majority of learning within ophthalmology is considered to be supervised because of its use of pre-labelled data during training. Deep learning is very resource intensive, and for this reason, the sizes of the input images are typically decreased during processing. Along these li
Data Loading...
