Probabilistic Principal Component Analysis and Particle Filtering for real-time retina detection from a single-fiber OCT

Vitreo-retinal surgery concerns a set of particularly demanding micro-surgical interventions that take place at the back of the eye. Examples of such procedures are retinal vein cannulation (where the surgeon aims to insert a needle in a vein of the size of human hairs) and epiretinal membrane peeling (where a detached membrane must be removed from the retina). As severe retinal damage can be caused by undesired collisions, good instrument to retina distance perception would be very useful. We propose to use an OCT-fiber instrumented tool, and purposefully designed algorithms to interpret the measurements and extract a reliable real-time distance estimate. This abstract describes the progress that was made and includes a test conducted with a robotic platform on a synthetic eye mockup.