11 Segmentation and Visualization of Drusen …
295
area. In the future, we could calculate other drusen features from these images, such
as shape.
11.2.2 An Improved OCT-Derived Fundus Projection Image
for Drusen Visualization
Currently, the gold standard for visualizing and measuring drusen in non-neovascular
AMD as well as for visualizing and assessing GA is the evaluation of color fundus
photographs (CFPs). While the total drusen area and maximum drusen size are estimated by visual inspection of CFPs, with comparison to a set of standard circles [39].
So, it is a big challenge to reliably locate drusen against the changing background
pigments of the macula, RPE, and choroid [40, 41]. Moreover, it is difficult to make
reproducible quantitative measurements of drusen in CFPs, and such measurements
could be better indicators of disease progression than qualitative visual assessments.
Since SD-OCT images provide 3D data, and the structures visualized in the volume can be projected into 2D, and the current method for creating 2D projections
from SD-OCT datasets is the summed-voxel projection (SVP), in which all pixel
values in the 3D images are summed along axial lines, producing an image showing
the retinal surface en face, similar to the CFP [36]. The SVP fundus image is not
good for drusen visualization because most drusen are not visible when projected
using this method [36]. Stopa [37] overcame some of these problems by locating
pathologic retinal features with color marking in each OCT image before the image
volume was collapsed along the depth axis to produce the SVP. Current technique
recently introduced into OCT imaging devices is the “slab SVP”, is a semi-automated
method to restrict the SVP to a sub-volume of the retina in vicinity of the RPE layer
(Carl Zeiss Meditch, Inc., unpublished); in this method, user interaction to annotate
the image to localize the RPE is required. Using manually annotating pathologic
features in a stack of SD-OCT images, for large studies, can be time-consuming and
tiring. More so, the SVP image produced by the proposed methods of Stopa et al.
only gives location information, but no information about drusen thickness, which
is useful for characterizing drusen. Georczynska [38] proposed a better method of
generating projection OCT fundus images by selectively summing different retinal
depth levels, which enhanced contrast and visualized outer retinal pathology not visible with standard fundus imaging or OCT fundus imaging. In this method, drusen
were separated into several projected fundus images summed at different retinal
depth levels, and could not be directly visualized.
In this section, we analyze the reasons for poor drusen visualization in SVP
fundus images, and present a new, automated projection method combined with
image processing of drusen to generate en face fundus images from SD-OCT for
enhanced drusen visualization [35].
295
area. In the future, we could calculate other drusen features from these images, such
as shape.
11.2.2 An Improved OCT-Derived Fundus Projection Image
for Drusen Visualization
Currently, the gold standard for visualizing and measuring drusen in non-neovascular
AMD as well as for visualizing and assessing GA is the evaluation of color fundus
photographs (CFPs). While the total drusen area and maximum drusen size are estimated by visual inspection of CFPs, with comparison to a set of standard circles [39].
So, it is a big challenge to reliably locate drusen against the changing background
pigments of the macula, RPE, and choroid [40, 41]. Moreover, it is difficult to make
reproducible quantitative measurements of drusen in CFPs, and such measurements
could be better indicators of disease progression than qualitative visual assessments.
Since SD-OCT images provide 3D data, and the structures visualized in the volume can be projected into 2D, and the current method for creating 2D projections
from SD-OCT datasets is the summed-voxel projection (SVP), in which all pixel
values in the 3D images are summed along axial lines, producing an image showing
the retinal surface en face, similar to the CFP [36]. The SVP fundus image is not
good for drusen visualization because most drusen are not visible when projected
using this method [36]. Stopa [37] overcame some of these problems by locating
pathologic retinal features with color marking in each OCT image before the image
volume was collapsed along the depth axis to produce the SVP. Current technique
recently introduced into OCT imaging devices is the “slab SVP”, is a semi-automated
method to restrict the SVP to a sub-volume of the retina in vicinity of the RPE layer
(Carl Zeiss Meditch, Inc., unpublished); in this method, user interaction to annotate
the image to localize the RPE is required. Using manually annotating pathologic
features in a stack of SD-OCT images, for large studies, can be time-consuming and
tiring. More so, the SVP image produced by the proposed methods of Stopa et al.
only gives location information, but no information about drusen thickness, which
is useful for characterizing drusen. Georczynska [38] proposed a better method of
generating projection OCT fundus images by selectively summing different retinal
depth levels, which enhanced contrast and visualized outer retinal pathology not visible with standard fundus imaging or OCT fundus imaging. In this method, drusen
were separated into several projected fundus images summed at different retinal
depth levels, and could not be directly visualized.
In this section, we analyze the reasons for poor drusen visualization in SVP
fundus images, and present a new, automated projection method combined with
image processing of drusen to generate en face fundus images from SD-OCT for
enhanced drusen visualization [35].