Automated drusen segmentation and quantification in SD-OCT images

Chen, Qiang; Leng, Theodore; Zheng, Luoluo; Kutzscher, Lauren; Ma, Jeffrey; Sisternes, Luís de; Rubin, Daniel L.

doi:10.1016/j.media.2013.06.003

Cited by 115 publications

(79 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For evaluation of drusen areas, we computed the overlap ratio (same as in [13]) and the absolute area difference (AAD). The ADD is given by AAD = |A man − A auto |, where A man and A auto are the drusen areas obtained from the manual segmentation and the segmentation of an automated method, respectively.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-surface segmentation of OCT images with AMD using sparse high order potentials

Oliveira

Pereira

Gonçalves³

et al. 2016

Biomed. Opt. Express

View full text Add to dashboard Cite

In age-related macular degeneration (AMD), the quantification of drusen is important because it is correlated with the evolution of the disease to an advanced stage. Therefore, we propose an algorithm based on a multi-surface framework for the segmentation of the limiting boundaries of drusen: the inner boundary of the retinal pigment epithelium + drusen complex (IRPEDC) and the Bruch's membrane (BM). Several segmentation methods have been considerably successful in segmenting retinal layers of healthy retinas in optical coherence tomography (OCT) images. These methods are successful because they incorporate prior information and regularization. Nonetheless, these factors tend to hinder the segmentation for diseased retinas. The proposed algorithm takes into account the presence of drusen and geographic atrophy (GA) related to AMD by excluding prior information and regularization just valid for healthy regions. However, even with this algorithm, prior information and regularization still cause the oversmoothing of drusen in some locations. Thus, we propose the integration of local shape prior in the form of a sparse high order potentials (SHOPs) into the algorithm to reduce the oversmoothing of drusen. The proposed algorithm was evaluated in a public database. The mean unsigned errors, relative to the average of two experts, for the inner limiting membrane (ILM), IRPEDC and BM were 2.94±2.69, 5.53±5.66 and 4.00±4.00 µm, respectively. Drusen areas measurements were evaluated, relative to the average of two expert graders, by the mean absolute area difference and overlap ratio, which were 1579.7 ± 2106.8 µm 2 and 0.78 ± 0.11, respectively.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Algorithms such as [10,12,13] use a similar structure, i.e., ILM segmentation, coarse and fine segmentation of one of the RPE boundaries and BM estimation through smoothing. These algorithms also use similar methods, such as thresholding, gradient peak detection and polynomial fitting.…”

Section: Introductionmentioning

confidence: 99%

Multi-surface segmentation of OCT images with AMD using sparse high order potentials

Oliveira

Pereira

Gonçalves³

et al. 2016

Biomed. Opt. Express

View full text Add to dashboard Cite

show abstract

“…The basic idea of most image denoising methods (such as bilateral filtering and HS-based method) is to find similar The location of the RPE boundary was automatically detected using a segmentation tool developed by our group that analyzes individual B-scan pixel statistics [30,31]. Figure 2b is the flattened image version of the original B-scan displayed in Fig.…”

Section: Neighborhood Shapementioning

confidence: 99%

Application of Improved Homogeneity Similarity-Based Denoising in Optical Coherence Tomography Retinal Images

et al. 2014

Self Cite

View full text Add to dashboard Cite

Image denoising is a fundamental preprocessing step of image processing in many applications developed for optical coherence tomography (OCT) retinal imaging-a highresolution modality for evaluating disease in the eye. To make a homogeneity similarity-based image denoising method more suitable for OCT image removal, we improve it by considering the noise and retinal characteristics of OCT images in two respects: (1) median filtering preprocessing is used to make the noise distribution of OCT images more suitable for patchbased methods; (2) a rectangle neighborhood and region restriction are adopted to accommodate the horizontal stretching of retinal structures when observed in OCT images. As a performance measurement of the proposed technique, we tested the method on real and synthetic noisy retinal OCT images and compared the results with other well-known spatial denoising methods, including bilateral filtering, five partial differential equation (PDE)-based methods, and three patchbased methods. Our results indicate that our proposed method seems suitable for retinal OCT imaging denoising, and that, in general, patch-based methods can achieve better visual denoising results than point-based methods in this type of imaging, because the image patch can better represent the structured information in the images than a single pixel. However, the time complexity of the patch-based methods is substantially higher than that of the others.

show abstract

“…The ability of SD-OCT to rapidly acquire three-dimensional (3D) retinal data, resolving structures in the depth axis, allows the direct visualization of its layered structure. Many new quantitative features extracted from SD-OCT data, such as the thickness of individual intraretinal layers [2], size, shape, and distribution of drusen (extracellular material accumulations that typically appear between the retinal pigment epithelium (RPE) and Bruch's membrane) [3,4], cysts and fluid-filled regions [5], and macular [6], are currently used or investigated as biomarkers in retinal disease diagnosis [7][8][9]. The accurate and reliable segmentation of retinal layers in SD-OCT scans is a fundamental problem for the identification of new quantitative features that could be useful as disease biomarkers.…”

Section: Introductionmentioning

confidence: 99%

Automated intraretinal segmentation of SD-OCT images in normal and age-related macular degeneration eyes

Sisternes

Jonna

Moss

et al. 2017

Biomed. Opt. Express

Self Cite

View full text Add to dashboard Cite

This work introduces and evaluates an automated intra-retinal segmentation method for spectral-domain optical coherence (SD-OCT) retinal images. While quantitative assessment of retinal features in SD-OCT data is important, manual segmentation is extremely time-consuming and subjective. We address challenges that have hindered prior automated methods, including poor performance with diseased retinas relative to healthy retinas, and data smoothing that obscures image features such as small retinal drusen. Our novel segmentation approach is based on the iterative adaptation of a weighted median process, wherein a three-dimensional weighting function is defined according to image intensity and gradient properties, and a set of smoothness constraints and pre-defined rules are considered. We compared the segmentation results for 9 segmented outlines associated with intra-retinal boundaries to those drawn by hand by two retinal specialists and to those produced by an independent state-of-the-art automated software tool in a set of 42 clinical images (from 14 patients). These images were obtained with a Zeiss Cirrus SD-OCT system, including healthy, early or intermediate AMD, and advanced AMD eyes. As a qualitative evaluation of accuracy, a highly experienced third independent reader blindly rated the quality of the outlines produced by each method. The accuracy and image detail of our method was superior in healthy and early or intermediate AMD eyes (98.15% and 97.78% of results not needing substantial editing) to the automated method we compared against. While the performance was not as good in advanced AMD (68.89%), it was still better than the manual outlines or the comparison method (which failed in such cases). We also tested our method's performance on images acquired with a different SD-OCT manufacturer, collected from a large publicly available data set (114 healthy and 255 AMD eyes), and compared the data quantitatively to reference standard markings of the internal limiting membrane and inner boundary of retinal pigment epithelium, producing a mean unsigned positioning error of 6.04 ± 7.83µm (mean under 2 pixels). Our automated method should be applicable to data from different OCT manufacturers and offers detailed layer segmentations in healthy and AMD eyes.

show abstract

Automated drusen segmentation and quantification in SD-OCT images

Cited by 115 publications

References 38 publications

Multi-surface segmentation of OCT images with AMD using sparse high order potentials

Multi-surface segmentation of OCT images with AMD using sparse high order potentials

Application of Improved Homogeneity Similarity-Based Denoising in Optical Coherence Tomography Retinal Images

Automated intraretinal segmentation of SD-OCT images in normal and age-related macular degeneration eyes

Contact Info

Product

Resources

About