Polarization-Based Eye Tracking with Personalized Siamese Architectures

2026-03-26

AI Summary

This study addresses the challenge of extensive calibration requirements in head-mounted eye tracking, which stem from individual anatomical differences and hinder natural human-computer interaction. To mitigate this burden, the work introduces polarimetric imaging into personalized gaze estimation for the first time, proposing a framework that integrates a polarization-sensitive camera under 850 nm illumination with a Siamese neural network. The model learns relative gaze displacements from only a few calibration frames to reconstruct absolute gaze points. Remarkably, the approach achieves performance comparable to conventional linear calibration using merely one-tenth of the typical calibration samples. Furthermore, polarization cues reduce gaze estimation error by up to 12% compared to standard near-infrared inputs, and when combined with linear calibration, yield an additional 13% accuracy improvement, substantially alleviating the calibration overhead.

Abstract

Head-mounted devices integrated with eye tracking promise a solution for natural human-computer interaction. However, they typically require per-user calibration for optimal performance due to inter-person variability. A differential personalization approach using Siamese architectures learns relative gaze displacements and reconstructs absolute gaze from a small set of calibration frames. In this paper, we benchmark Siamese personalization on polarization-enabled eye tracking. For benchmarking, we use a 338-subject dataset captured with a polarization-sensitive camera and 850 nm illumination. We achieve performance comparable to linear calibration with 10-fold fewer samples. Using polarization inputs for Siamese personalization reduces gaze error by up to 12% compared to near-infrared (NIR)-based inputs. Combining Siamese personalization with linear calibration yields further improvements of up to 13% over a linearly calibrated baseline. These results establish Siamese personalization as a practical approach enabling accurate eye tracking.

Problem

• eye tracking
• personalization
• calibration
• polarization
• Siamese architectures

Innovation

• polarization-based eye tracking
• Siamese architecture
• personalized calibration
• gaze estimation
• head-mounted eye tracking