Automated construction of arterial and venous trees in retinal images

Abstract

Abstract. While many approaches exist to segment retinal vessels in fundus photographs, only a limited number focus on the construction and disambiguation of arterial and venous trees. Previous approaches are local and/or greedy in nature, making them susceptible to errors or limiting their applicability to large vessels. We propose a more global framework to generate arteriovenous trees in retinal images, given a vessel segmentation. In particular, our approach consists of three stages. The first stage is to generate an overconnected vessel network, named the vessel potential connectivity map (VPCM), consisting of vessel segments and the potential connectivity between them. The second stage is to disambiguate the VPCM into multiple anatomical trees, using a graph-based metaheuristic algorithm. The third stage is to classify these trees into arterial or venous (A/V) trees. We evaluated our approach with a ground truth built based on a public database, showing a pixelwise classification accuracy of 88.15% using a manual vessel segmentation as input, and 86.11% using an automatic vessel segmentation as input.

Artery/vein classification using reflection features in retina fundus images

“…Therefore, the separation of arteries and veins on the retinal images provides new information apart from the usual biomarkers. Several works on separating retinal arteries and veins have been proposed in the literature [9][10][11][12][13][14][15]. From the recent works, we can summarize the workflow for an artery/vein (A/V) classification system.…”

“…The label of each vessel centerline, or vessel segment, is determined by averaging the labels of its pixels, called the segment-wise classification [11,14]. Finally, the result of segment-wise classification is corrected by using contextual information extracted from the vascular structure, like vessels connecting with each other have the same label and in crossovers they have opposite labels [9,10,12].…”

“…During the last few years, the application of graph-based approaches to classify arteries and veins has become popular in retinal image analysis [9][10][11][12]. A graph of the vascular tree represents the topological structure of the vessels.…”

“…In addition, Estrada et al [12] obtain the local RGB intensity and compute the mean color value of the three channels. The framework proposed by Hu et al [10] uses a similar method by Niemeijer et al [13] where the feature vector is based on the red, green, hue, saturation and brightness intensities, which are normalized to zero mean and unit standard deviation.…”

A fluid-dynamic based approach to reconnect the retinal vessels in fundus photography

“…An efficient search algorithm explores the solutions space and eventually provides the vascular topology and the final artery and vein labeled trees. Hu et al [12] proposed a global three-step framework. It starts with the generation of an over-connected vessels network.…”