Loosely coupled level sets for retinal layers and drusen segmentation in subjects with dry age-related macular degeneration

Abstract

Important noteTo cite this publication, please use the final published version (if applicable). Please check the document version above. CopyrightOther than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons. Takedown policyPlease contact us and provide details if you believe this document breaches copyrights. We will remove access to the work immediately and investigate your claim. ABSTRACTOptical coherence tomography (OCT) is used to produce high-resolution three-dimensional images of the retina, which permit the investigation of retinal irregularities. In dry age-related macular degeneration (AMD), a chronic eye disease that causes central vision loss, disruptions such as drusen and changes in retinal layer thicknesses occur which could be used as biomarkers for disease monitoring and diagnosis. Due to the topology disrupting pathology, existing segmentation methods often fail. Here, we present a solution for the segmentation of retinal layers in dry AMD subjects by extending our previously presented loosely coupled level sets framework which operates on attenuation coefficients. In eyes affected by AMD, Bruch's membrane becomes visible only below the drusen and our segmentation framework is adapted to delineate such a partially discernible interface. Furthermore, the initialization stage, which tentatively segments five interfaces, is modified to accommodate the appearance of drusen. This stage is based on Dijkstra's algorithm and combines prior knowledge on the shape of the interface, gradient and attenuation coefficient in the newly proposed cost function. This prior knowledge is incorporated by varying the weights for horizontal, diagonal and vertical edges. Finally, quantitative evaluation of the accuracy shows a good agreement between manual and automated segmentation.

Active contour method for ILM segmentation in ONH volume scans in retinal OCT

“…Novosel et al [30] proposed a 3D approach to jointly segment retinal layers and lesions in eyes with topology-disrupting retinal diseases that can handle local intensity variations as well as the presence or absence of pathological structures in the retina. This method extends and generalizes the existing framework of loosely coupled level set, previously developed for the simultaneous segmentation of interfaces between retinal layers in macular and peripapillary scans of healthy and glaucoma subjects [31], drusen segmentation in subjects with dry age-related macular degeneration [32] and fluid segmentation in subjects with central serous rethinopathy [33]. Some of these methods are still 2D based, [32,33], and all mainly focus on the macula.…”

“…This method extends and generalizes the existing framework of loosely coupled level set, previously developed for the simultaneous segmentation of interfaces between retinal layers in macular and peripapillary scans of healthy and glaucoma subjects [31], drusen segmentation in subjects with dry age-related macular degeneration [32] and fluid segmentation in subjects with central serous rethinopathy [33]. Some of these methods are still 2D based, [32,33], and all mainly focus on the macula. Carass et al [34] proposed a 3D level set approach based on multiple-object geometric deformable model that overcomes the voxel precision restrictions that are present in the aforementioned methods.…”

“…Several level-sets methods have also been employed in retinal layer segmentation, most of which are related to the macula region and, or specific pathologies in this region [30][31][32][33][34]. [32,33] use 2D segmentation, which limits their capabilities in capturing challenging 3D structures, as seen at the ONH. Furthermore, most of these methods focus on the macula, except [31] which also provides data on ONH scans, although as the authors stated themselves, this algorithm would need further modifications to account for difficult topologies.…”

Automated retinal layer segmentation in OCT images of age‐related macular degeneration

“…Next, for segmenting ILM and RPEDC layer boundaries, we use the dark-to-light vertical gradient to compute edge weight w 1 (a, b) according to Formula (7) where I dla and I dlb are the dark-to-light vertical gradients at nodes a and b, respectively, i a and i b are the intensity value at nodes a and b, respectively, λ s = 0.1 is a constant, w min = 1 × 10 −5 is the minimum constant which increases the stability of the algorithm, '0' indicates the nodes a and b are disconnected. To segment BM layer boundary, we use the light-to-dark vertical gradient to compute edge weights w 2 (a, b) according to formula (8). Where I lda and I ldb are the light-to-dark vertical gradients at nodes a and b, respectively, y a and y b are the y coordinates of nodes a and b, respectively, w v = 1 is a constant to keep the BM layer boundary flat, '0' indicates the nodes a and b are disconnected.…”

“…Novosel et al [7] developed a segmentation method that used loosely coupled level sets to segment eight retinal layers. This approach was later extended for segmentation of retinal layer in dry AMD subjects [8]. Grapy theory and dynamic programming (GTDP) method was used by Chiu et al [9] to segment eight retinal layers of normal subjects.…”