Automatická analýza a modelování terče zrakového nervu z retinálních záznamů

Abstract

This master thesis is devoted to the theme of automatic segmentation and modeling of optic disc from retinal images. The proposed algorithm consists of several steps. First step of the algorithm is image pre-processing. Second part of algorithm is devoted to the image segmentation. For the image segmentation, this master thesis proposed two algorithms. First image segmentation algorithm is fully automatic and uses a swarm intelligence algorithm. The second image segmentation algorithm uses an assisted segmentaion and is determinated for images with low resolution. For the optic disc boundary detection the algorithm uses an active contour method. The last step of algorithm is de-voted to the extraction of optic disc geometrical features. Thesis contains a brief introduction to the illnes rethinopathy of prematurity, possibilities of diagnostic process, medical treatment and scree-ning. Next topic is devoted to the digital screening system RetCam3. Last part of theoretical text was devoted to the embryonic development of human eye. Master thesis contains a complex re-aserch of the newest methods and trends in image segmentation and optic disc detection from reti-nal images.