%0 Journal Article
%A Moura Ramos, José Joaquim, de
%A Fernández Vigo, José Ignacio
%A Martínez de la Casa, Jose Maria
%A García Feijoo, Julián
%A Gende Lozano, Mateo
%A Novo Buján, Jorge
%A Ortega Hortas, Marcos
%T Robust multi-view approaches for retinal layer segmentation inglaucoma patients via transfer learning
%D 2023
%@ 2223-4292
%U https://hdl.handle.net/20.500.14352/73120
%X Background: Glaucoma is the leading global cause of irreversible blindness. Glaucoma patients experiencea progressive deterioration of the retinal nervous tissues that begins with a loss of peripheral vision. An earlydiagnosis is essential in order to prevent blindness. Ophthalmologists measure the deterioration caused bythis disease by assessing the retinal layers in different regions of the eye, using different optical coherencetomography (OCT) scanning patterns to extract images, generating different views from multiple parts ofthe retina. These images are used to measure the thickness of the retinal layers in different regions.Methods: We present two approaches for the multi-region segmentation of the retinal layers in OCTimages of glaucoma patients. These approaches can extract the relevant anatomical structures for glaucomaassessment from three different OCT scan patterns: circumpapillary circle scans, macular cube scans andoptic disc (OD) radial scans. By employing transfer learning to take advantage of the visual patterns presentin a related domain, these approaches use state-of-the-art segmentation modules to achieve a robust, fullyautomatic segmentation of the retinal layers. The first approach exploits inter-view similarities by using asingle module to segment all of the scan patterns, considering them as a single domain. The second approachuses view-specific modules for the segmentation of each scan pattern, automatically detecting the suitablemodule to analyse each image.Results: The proposed approaches produced satisfactory results with the first approach achieving a dicecoefficient of 0.85±0.06 and the second one 0.87±0.08 for all segmented layers. The first approach producedthe best results for the radial scans. Concurrently, the view-specific second approach achieved the best resultsfor the better represented circle and cube scan patterns.Conclusions: To the extent of our knowledge, this is the first proposal in the literature for the multi-viewsegmentation of the retinal layers of glaucoma patients, demonstrating the applicability of machine learningbased systems for aiding in the diagnosis of this relevant pathology.
%~