6533b7ddfe1ef96bd12746a5

RESEARCH PRODUCT

Microaneurysms detection with the radon cliff operator in retinal fundus images

subject

description

ABSTRACT Diabetic Retinopathy (DR) is one of the leading causes of blindness in the industrialized world. Early detection is thekey in providing effective treatment. However, the current number of trained eye care specialists is inadequate to screenthe increasing number of diabetic patients. In recent years, automated and semi-automated systems to detect DR withcolor fundus images have been developedwith encouraging,but not fully satisfactory results. In this study we present theinitial results of a new techniquefor the detection and localization of microaneurysms,an early sign of DR. The algorithmis based on three steps: candidates selection, the actual microaneurysms detection and a Þnal probability evaluation. Weintroducethenew RadonCliff operatorwhichisourmaincontributiontotheÞeld. MakinguseoftheRadontransform,theoperator is able to detect single noisy Gaussian-like circular structures regardless of their size or strength. The advantagesoverexistingmicroaneurysmsdetectorsaremanifold: thesizeofthelesionscanbeunknown,itautomaticallydistinguisheslesions from the vasculature and it provides a fair approach to microaneurysm localization even without post-processingthe candidates with machine learning techniques, facilitating the training phase. The algorithm is evaluated on a publiclyavailable dataset from the Retinopathy Online Challenge.Keywords: radon transform,diabetic retinopathy,segmentation