Premium
Trial study to automatically distinguish small haemorrhages in early diabetic retinopathy from image artefacts
Author(s) -
Suzuki N.,
Yamane K.
Publication year - 2016
Publication title -
acta ophthalmologica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.534
H-Index - 87
eISSN - 1755-3768
pISSN - 1755-375X
DOI - 10.1111/j.1755-3768.2016.0425
Subject(s) - rgb color model , fundus (uterus) , cotton wool spots , fundus photography , artificial intelligence , diabetic retinopathy , computer science , computer vision , mathematics , medicine , ophthalmology , diabetes mellitus , retinal , fluorescein angiography , endocrinology
Purpose The L*u*v* colour space presented optimal results, with the highest sensitivity and best reproducibility, among RGB, XYZ, CMY, HSL, HSV, HSI and L*a*b* colour spaces. Therefore, we employed three‐dimensional analysis of L*u*v* colour spaces to detect early diabetic retinopathy. Methods Six patients with small haemorrhages were evaluated using fundus photography, which revealed image artefacts in the fundi of some patients. We constructed an experimental device similar to the optical system of a fundus camera and created artificial eyes of the fundus, which were painted with five different colours: rose, coffee, red, orange and yellow. The image artefacts were photographed under each artificial eye using the experimental device. In addition, the following eight types of specimen were used: a dust particle, wool, a piece of paper, a wood chip, cotton, a grey hair, a drop of water and a piece of plastic bag. All images were analysed using Scilab 5.4.0 and SIVP 0.5.3 softwares. Results We constructed an algorithm to calculate the difference between the averages of the central and circumference areas. In all image artefacts, L*u*v* colour spaces was highly sensitive: L* values were 2.8–8.5, u* values were 3.8–21 and v* values were 4.2–10.1. Conclusions We succeeded in automatically distinguishing small haemorrhages in early diabetic retinopathy from image artefacts.