Original title:
Sémantická segmentace patologií v obrazech sítnice
Translated title:
Semantic Segmentation of Pathologies in Retinal Images
Authors:
Čabala, Roman ; Orság, Filip (referee) ; Kavetskyi, Andrii (advisor) Document type: Master’s theses
Year:
2023
Language:
slo Publisher:
Vysoké učení technické v Brně. Fakulta informačních technologií Abstract:
[slo][eng]
Cieľom diplomovej práce bolo segmentovať patológiu viditeľnú na snímkach sietnice, ako sú exsudáty, hemoragia a mikroaneuryzmy. Za týmto účelom boli vyskúšané dve dobre známe hlboké neurónové siete, konkrétne U-Net a SegFormer. Na testovanie výkonnosti modelov sa použil jeden verejne dostupný dataset IDRiD. Získané výsledky boli opísané po analýze rôznych faktorov, ktoré ovplyvnili výkon modelov U-Net a Segformer.
The thesis aimed to segment pathology visible in the retina images, such as exudates, hemorrhages, and microaneurysms. For that, two well known deep neural networks, named U-Net and SegFormer, were trained. To test the performance of the models, one publicly available dataset was used, named IDRiD. Obtained results were reported after analyzing different factors which affected the performance of the models U-Net and Segformer.
Keywords:
drusen; exudates; hemorrhages; IDRiD; microaneurysms; retina; SegFormer; segmentation transformers; semantic segmentation; SETR; U-Net
Institution: Brno University of Technology
(web)
Document availability information: Fulltext is available in the Brno University of Technology Digital Library. Original record: http://hdl.handle.net/11012/213212