Establishing Robust Retinal Eye Tracking: A Weakly Supervised Algorithmic Framework

TL;DR

This paper introduces a weakly-supervised, learning-based framework for retinal eye tracking that improves robustness and accuracy over traditional methods, achieving a 95th-percentile gaze error below 0.45 degrees.

Contribution

The work presents a novel weakly-supervised algorithmic framework that enhances robustness and accuracy in retinal eye tracking under real-world conditions.

Findings

Achieved 95th-percentile gaze error < 0.45 degrees across 6 participants.

Demonstrated high accuracy with the proposed learning-based approach.

Outperformed classical template-matching methods in robustness.

Abstract

Retinal image-based eye tracking is widely used in ophthalmic imaging and vision science, and is a promising path to deliver higher gaze accuracy than the pupil- and cornea-based approaches commonly used in modern AR/VR devices. Nevertheless, existing retinal tracking algorithms still primarily rely on classical template-matching registration, which can be insufficiently robust to retinal feature variability and real-world imaging conditions. In this work, we propose a novel weakly-supervised, learning-based framework for robust retinal eye tracking. Initial studies demonstrate high accuracy, achieving the 95th-percentile gaze error < 0.45 deg across a cohort of 6 participants.