Purchase this article with an account.
Hansford Hendargo, Rolando Estrada, Stephanie Chiu, Carlo Tomasi, Sina Farsiu, Joseph Izatt; Image Registration for Motion Artifact Removal in Retinal Vascular Imaging Using Speckle Variance Fourier Domain Optical Coherence Tomography. Invest. Ophthalmol. Vis. Sci. 2013;54(15):5528.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Variance processing methods in Fourier domain optical coherence tomography (FD-OCT) have enabled depth-resolved visualization of the capillary beds in the retina. However, acquisition of variance datasets requires several seconds, even with high speed systems. Eye motion during this time span is sufficient to corrupt visualization of the vasculature. We demonstrate a method to eliminate motion artifacts in speckle variance FD-OCT images of the retinal vasculature by creating a composite image from multiple volumes of data.
Volume datasets collected from subjects with a swept-source retinal OCT system (1060 nm, 100 kHz) consisted of 300 A-scans/B-scan, 3 repeated B-scans, and 300 B-scans locations over a 2.5x2.5 mm area centered at the fovea. Volumes with fast scanning along either the x- or y-axis were acquired. B-scans were axially registered, and the intensity variance over each of the 3 repeated frames was computed. Automated segmentation of the retinal layers was used to create summed volume projections (SVPs) from each retinal layer. Motion artifacts were detected using an intensity threshold to divide each SVP into motion-free strips. Each strip was Gabor filtered, and strips from multiple volumes were globally registered via X-Y translation and locally registered using a free form deformation method based on b-splines. Registration was performed using the SVPs generated from the IPL, and the image transform parameters were applied to subsequent layers.
Figure 1 shows representative data collected from a healthy 27 year-old volunteer. Individual SVPs show motion artifact streaks, and the registered images appear motion-free. Vessels appear to be smoothly connected, and visualization of the foveal avascular zone is enhanced in each of the vessel layers.
We demonstrate the ability to remove motion artifacts through processing methods for visualization of the distinct capillary beds. Variance OCT datasets may reveal clinically useful information on pathologic effects specific to the vasculature of the different retinal layers.
This PDF is available to Subscribers Only