Abstract
Purpose :
Robotically Aligned OCT (RAOCT) systems can image patients without a fixation target or stabilization (chin/forehead rest). Additionally, through robotic angular positioning control, RAOCT enables automatic acquisition of retinal images at different regions of interest (ROI), which can subsequently be montaged to synthetically increase the field-of-view (FOV) of the system.
Methods :
Utilizing custom software, pupil tracking cameras estimated ocular pupil location and gaze of the eye, enabling autonomous alignment of a robotic arm mounted 20° FOV SSOCT retinal scanner (1050nm; Fig. 1A). The tracked gaze was utilized to control the orientation of the scanner relative to the optic axis of the eye, allowing control over the OCT beam’s angle of incidence into the eye. Control of the angle of incidence allowed imaging multiple ROI by pivoting the scanner about the pupil (Fig. 1B). For each imaging session, the robot automatically aimed at and imaged 25 different ROIs over a 20° range. We subsequently montaged acquired volumes onto a single large FOV volume utilizing the gaze tracking information for localization and further optimized using b-spline based registration. We imaged consented freestanding subjects under an IRB approved protocol.
Results :
Imaging performance on a healthy 26-year-old is demonstrated in Figure 2. Retinal montaging allowed us to synthetically double the FOV of the system from 20° to 40° with automatic alignment and aiming of the OCT scanner with minimal discontinuities along the retinal vasculature. The mean residual error corrected with registration was 0.21° across the entire field of view.
Conclusions :
We demonstrated automatic retinal volumetric acquisition over multiple ROIs over a 20° range in freestanding subjects. With montaging, we demonstrated increased image FOV comparable to fundus photography.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.