P139 Automatic Classification of Arterial and Venular Trees in Colour Fundus Images

Abstract

Background

Quantitative imaging of retinal arterioles and venules offers unique insights into cardiovascular and microvascular diseases but is laborious. We developed and tested a method to automatically identify Arte-rial/Venular (A/V) vessels in digital retinal images in conjunction with a semi-automatic segmentation technique.

Methods

Segmentation of blood vessels and the Optic Disc (OD) was performed as previously described [1] using a dataset of X colour fundus images. Using the OD as a reference point a graph representation was constructed using the vessel skeletons. Vessel bifurcations and crossings were identified based on direction and local geometry, and A/V classification was carried out by fuzzy logic classification using colour information. Results were compared with expert classification.

Results

157 arterial and 150 venular segments were classified. Preliminary Results showed sensitivity, specificity and accuracy of 42.20%, 99.21% and 97.73% for arteries and 50.89%, 98.70% and 97.54% for veins. An example is shown in Figure 1.

Conclusions

Computer-based systems can assess local and global aspects of the retinal microvascular architecture, geometry and topology. Automated A/V classification will facilitate efficient cost-effective assessment of clinical images at scale.