Abstract
Purpose :
Early indications of some of the retinal diseases like ocular ischemic syndrome, retinal detachment and sickle cell retinopathy are apparent in the periphery of retina. Montaging fundus images gives us the ability to view the extreme periphery of the retina. Before montaging, the anchor point defined at the fovea, can be manually chosen by the user or automatically detected using an algorithm. The purpose of this study was to compare the image quality of ultra-widefield (UWF) montaged images generated using manual and automated anchor point finding.
Methods :
TrueColor widefield (WF) fundus images (50) were captured on 25 subjects per eye (19 healthy and 6 diseased eyes) using CLARUS™ 500 (ZEISS, Dublin, CA). The montaging workflow involved anchor point finding (APF), distortion correction, feature finding and montaging. Anchor point (fovea) selection was performed manually and automatically for comparison. The user executed the manual and automated APF on WF images and generated montaged ultra WF (UWF) images. For qualitative evaluation, montage images were graded by the clinical expert for image quality considering vessel breakage, uniformness of both halves, artifacts, and color balance on the scale of 1 to 5 (1: unsuccessful, 2: poor, 3: fair, 4: good, 5: excellent). The image quality grade results of montaged images for manual and automated process were compared. Statistical analysis was performed to assess differences in the image quality.
Results :
Both, manual and automated montage workflow received image quality grade average of 4.7 indicating good to excellent image quality. Paired t test (p=1) results do not show statistically significant difference with 95% confidence interval in the manual versus automated image quality.
Conclusions :
In this study, the image quality of the ultra-widefield montaged images generated using manual and automated anchor point finding were comparable. This shows that the performance of automated anchor point finding process is reliable in generating montaged images with clinically acceptable image quality.
This abstract was presented at the 2019 ARVO Imaging in the Eye Conference, held in Vancouver, Canada, April 26-27, 2019.