Diabetic Retinopathy Detection in Ocular Imaging by Dictionary Learning Publisher

Summary: A study shows dictionary learning enhances diabetic eye disease detection, simplifying diagnosis with imaging. #EyeHealth #DiabetesResearch

Amini Z¹ ; Kafieh R¹ ; Mousavi E¹ ; Rabbani H¹
Source: Diabetes and Fundus OCT Published:2020

Abstract

This chapter discusses the power of dictionary learning (DL) in ocular image modeling, with most emphasis on diabetic retinopathy (DR) detection. For this purpose, anatomical biomarkers in DR detection are firstly introduced in this chapter. To demonstrate the previous background on the automatic detection of DR, a mini-review is then presented on DR classification developed over the past 2 decades. DL is also elaborated in this chapter and due to its capabilities, classification based on DL modeling is described in detail. Considering the fact that two prevalent ocular imaging modalities in DR detection are fundus imaging and optical coherence tomography (OCT), review and methods are presented in two distinct subsections. To conclude, in the proposed classification method, the need for preprocessing, segmentations, and feature extraction stages are eliminated based on the ability of DL in the classification of each image, more perceptually. © 2020 Elsevier Inc. All rights reserved.