Abstract
Purpose: :
Fundus angiograms convey valuable information about retinal blood vessels and have been widely used in the diagnosis of blood vessel alterations. Traditional fundus angiograms usually only image two dimensions. Optical coherence tomography (OCT), a non–invasive imaging technique offering micrometer resolution and 3D imaging capabilities, has become a prominent biomedical imaging modality. 3D OCT images can provide not only the 3D information of the retina but also the 3D structure of pathologies and structures of interest through 3D information extraction. The aim of this study is to extract 3D structural information of retinal blood vessels from OCT images and compare the 3D blood vessel structures among normal and diseased eyes.
Methods: :
3D OCT data set of size 1365 (depth, Z direction) × 512 (horizontal, X direction) × 128 (vertical, Y direction) covering a retina area of 2mm × 6mm × 6mm was acquired from a spectral–domain OCT system. A fundus intensity image was constructed from the same OCT data set. Image segmentation of blood vessels was implemented in two steps: firstly, the retinal blood vessel pattern in X–Y directions was acquired from the fundus image. Secondly, the blood vessel information in Z direction was obtained by blood vessel segmentation of each cross–sectional B–scan image. Reconstruction of 3D fundus angiogram from these blood vessel segmentation results has been executed by image processing method of volume rendering.
Results: :
3D fundus angiograms for normal subjects and eye diseased patients have been achieved from sets of OCT cross–sectional B–scan images. Transverse, sagittal, coronal, and custom views of the 3D angiogram have been obtained and saved. Comparison of the 3D retinal blood vessel structures of normal and diseased eyes has been implemented.
Conclusions: :
OCT B–scan images contain rich information about blood vessels. Visualization of a 3D fundus angiogram from OCT images reveals these details and provides a tool for further analysis of retinal vascular anatomy in relation to other structures of interest.
Keywords: image processing • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • blood supply