Purchase this article with an account.
Yiheng Lim, Roy de Kinkelder, Barry Cense; Cost-Effective Active Retinal Tracker to Stabilize Eye Motions in Optical Coherence Tomography. Invest. Ophthalmol. Vis. Sci. 2013;54(15):1499. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Functional imaging with optical coherence tomography (OCT) requires a stable imaging platform with active retinal tracking. A cost effective and robust design is paramount for successful application in the clinic.
A line-scanning laser ophthalmoscope (LSLO) as a retinal motion detector was integrated into a research-grade OCT system at 840 nm. The LSLO only adds a few extra optical components to the system, such as beam splitters, lenses and a linescan camera, thereby keeping the active retinal tracker simple and cost effective. Simultaneously with each OCT B-scan, a fundus image was captured with the LSLO. By cross-correlating LSLO fundus images, transverse eye motions were detected and an offset signal was sent to the OCT galvanometer to keep the field of view stable. The closed-loop speed of the tracking system was 8.7 Hz. In a tracking experiment, the left optic nerve head (ONH) of a healthy subject was imaged. As shown in Fig. 1, a 2 by 2 grid, corresponding to an area of 4° by 4° on the retina, was used as a fixation target for the tracking experiment. Experiments were performed with and without retinal tracking. As indicated by arrows in Fig. 1, starting at the center of the grid, the subject was asked to shift fixation every one second and follow nodes in a counter clockwise manner.
Fig. 2 shows OCT images with and without retinal tracking, corresponding to the positions in Fig. 1 marked a-d. Without retinal tracking, the OCT images in Fig. 2 (left) showed different regions of the ONH as the retina of the subject shifted. When the retinal tracker was turned on, as shown in Fig. 2 (right), the OCT images are similar indicating that the tracker was able to lock onto the ONH. Using the cross-correlated LSLO images, the location of the OCT images deviate from the target location by standard deviation 0.49° (X) and 0.24° (Y).
Voluntary eye motions were successfully compensated and OCT scans at the same retinal position independent of eye motion were obtained.
This PDF is available to Subscribers Only