查看文章

Distinguishing vessels in a retinal image is a solid first step towards a computer-aided diagnosis of several medical conditions. Mathematical morphology provides simple yet powerful transformations to process images. Extended with fuzzy techniques, it achieves the ability to handle the uncertainty present in medical imaging, both in the images and in the expert diagnosis. Within this paradigm, we propose a method that leverages the geometric and photometric features that characterize vessels as such. It is tested with two databases, DRIVE and STARE. The first promising results outperform some of the state of the art methods with which it was compared, positioning it among the most competitive techniques.