AutoAnaVeDiRe - Semi-automatic analysis of vessel parameters as potential risk factor for development and manifestation of diabetic retinopathy


Diabetic retinopathy as one of the eventual consequences of Diabetes Mellitus can lead to severe impairment of vision or even blindness. An early enough detection as well as an optimal diagnostic interval of this disease is crucial for an effective therapy. Detection and diagnosis of Diabetic Retinopathy is performed on digital images of the fundus of the eye, very often obtained with a fundus camera in red free mode. Currently, these images are manually evaluated by an ophthalmologist.
The different stages of diabetic retinopathy are connected with certain types of lesions which affect the small blood vessels in the retina. The retinal vessels can break down, cause leakages or get blocked. This affects the transport of oxygen and nutritive substances to the retina. More severe damage can develop when new blood vessels grow on the surface of the retina and leak fluid or bleed. All this changes are connected to the retinal vessel system which can possibly be used as an overall marker to predict the development of the disease.
Software tools are available to analyze the vessel diameter in fundus images (Siva, Vesselmap, Vampire). But to compare the results between several examinations over time a standard protocol has to be followed. This includes technical parameters like using the same type of fundus camera, field of view, color depth, brightness and location of analyzation. To investigate the retinal vessel system as a predictive marker new images need to be collected over several years regarding the standard protocol leaving already available fundus images unused.
The purpose of this project is to methodically find the restrictions and fitting solutions that would allow further usage of already existing cohort fundus images from long term studies. Old cohorts that have been followed up for more than 20 years, in a time of constant technical development often include images from multiple camera systems and imaging standards. This entails e.g. color or red-free imaging modes, different angle of view or macula or optic disk centered view points for the same patient. By statistically analyzing not just the changed vessel parameters analyzed under those circumstances but by also quantifying the image changes (e.g. resolution, transformation, quality) this project contributes new view points on new and finished studies using old images.


Für dieses Projekt wurden noch keine Projektpartner hinterlegt.


Für dieses Projekt wurden noch keine Förderungen hinterlegt.


  • Für dieses Projekt wurden noch keine Publikationen hinterlegt.