Artificial intelligence in glaucoma: opportunities, challenges, and future directions

Abstract

Artificial intelligence (AI) has shown excellent diagnostic performance in detecting various complex problems related to many areas of healthcare including ophthalmology. AI diagnostic systems developed from fundus images have become state-of-the-art tools in diagnosing retinal conditions and glaucoma as well as other ocular diseases. However, designing and implementing AI models using large imaging data is challenging. In this study, we review different machine learning (ML) and deep learning (DL) techniques applied to multiple modalities of retinal data, such as fundus images and visual fields for glaucoma detection, progression assessment, staging and so on. We summarize findings and provide several taxonomies to help the reader understand the evolution of conventional and emerging AI models in glaucoma. We discuss opportunities and challenges facing AI application in glaucoma and highlight some key themes from the existing literature that may help to explore future studies. Our goal in this systematic review is to help readers and researchers to understand critical aspects of AI related to glaucoma as well as determine the necessary steps and requirements for the successful development of AI models in glaucoma.

Keywords: Artificial intelligence; Deep learning; Glaucoma; Machine learning.

Fig. 1

Illustration and definition of artificial intelligence (AI), machine learning (ML), and deep learning (DL)

Fig. 2

Various types of ML models applied to glaucoma. GMM: Gaussian Mixture Modeling; PCA: Principal Component Analysis; NMF: Non-negative Matrix Factorization; AA: Archetypal Analysis; PCC: Pearson Correlation Coefficient, MB: Markov Blanket; mRMR: Minimum Redundancy Maximum Relevance

Fig. 3

Classification of DL models. CNN: Convolutional Neural Network; RNN: Recurrent Neural Network; LSTM: long short-term memory; DCGAN: Deep Convolutional Generative Adversarial Network; SSCNN: Convolutional Neural Network model with self-learning; SSCNN-DAE: Semi-supervised Convolutional Neural Network model with autoencoder

Fig. 4

Proposed review methodology for sample collection and analysis