A Simple Analytic Approach for Tracking Retinal Vessels and Measuring their Diameters

Retinal image processing provides tools for automatic diagnosis and monitoring of retinal diseases such as diabetic retinopathy (DR), age related macular degeneration (ARMD), glucoma and such. The properties of vessel structures on the other hand are widely utilized in locating morphologic structures such as optic disc and macula and in automatic diagnosis of the retinal diseases. Due to the importance of retinal vessels, we propose a simple approach for vessel tracking and measuring vessel diameter in retinal fundus images. Images having manually segmented retinal vasculatures are obtained from STARE database and used in this study. Our method first finds the midlines of the vessel network on the segmented images by employing Zhang-Suen thinning algorithm and then tracks the vessel branches through those midlines. Lastly, the diameters of the vessel segments in different parts of the vasculature are calculated along with the tracking operation. The performed test results show that the proposed automatic method is quite successfully tracks the vessel network and measure the diameter.