Premature infant blood vessel segmentation of retinal images based on hybrid method for the determination of tortuosity

Anotace: For the retinal blood vessels segmentation, we used a method, which is based on the morphological operations. The output of this process is extracted retinal binary image, where is situated main blood vessels. In this paper is used dataset of images (2800 images) from device RetCam3. Before applying the image processing, it was selected 30 images with diagnosed pre-plus diseases, and it is divided into two groups with low contrast and good contrast images. In the next part of the analysis, it was analyzing and showing blood vessels with tortuosity. Tortuosity is a symptom of ROP (retinopathy of prematurity). The clinical physicians evaluate tortuosity by visual comparison of the retinal images images. For this reason, it was suggested model which can automatically indicate the tortuosity of the retinal blood vessels by setting a threshold of the blood vessels curvature.