Supervised Retinal Vessel Segmentation Based on Neural Network Using Broader Aging Dataset

Retina image quality is affected by numerous factors including aging, refractive condition, and media opacity. Distracters that exist in certain age groups may not be present in another. This is evident when the retinal nerve fiber is more visible in younger age group, tricking the vessel segmentation algorithm to label it as vessel thus affecting the specificity performance of the supervised retinal vessel segmentation. This research work aims to investigate the impact of aging to the performance of the supervised vessel segmentation. The results suggest different age groups affect different aspect of the segmentation performance. Sensitivity is estimated to reduce by 4.633% for every 10year increase of age (p<0.001), and specificity is estimated to reduce by 0.543% for every 10-year decrease of age (p<0.001).