
Antoine Boniface
Office Manager, AlphaLum
Antoine Boniface spoke at MicroLED Connect in 2026, representing AlphaLum.
Camera-Free, Low-Power Eye Sensing for AI Glasses Using Self-Mixing Interferometry
Talk · 17 September 2026
→ See the MicroLED Connect & AR/VR Connect 2026 agenda
Accurate, high speed estimation of eye position and motion, through absolute gaze angle and its dynamics, is a high-value feature to enable natural user interfaces for AR and AI glasses. However, achieving high precision and frame rates under tight power constraints remains challenging for current camera based approaches, hindering their integration in all-day-wearable glasses. To address these limitations, we present an always on, camera free gaze sensing architecture based on self mixing interferometry (SMI). Unlike conventional methods that rely on spatial imaging, our approach uses temporal signals to capture rich eye movement information at rates up to 1000 Hz with ultra low power consumption. A volume phase hologram (VPH) integrated into the lens serves as the system’s sole optical element, redirecting light toward the eye and enabling detection of minute eye displacements over a wide eye box. With an appropriate machine learning pipeline, the proposed system supports eye tracking using multiple sensors and, from a single SMI sensor, can detect a range of eye movement primitives, including fixations, saccades, micro-saccades, blinks, catch up saccades, and smooth pursuit, while estimating their kinematics.
▶ Watch video highlights from this talk
- ▶ How can a single VCSEL and photodiode replace a camera for low-power eye tracking?
- ▶ Can a glasses platform classify eye primitives such as saccades and smooth pursuits with high accuracy?
- ▶ What is the power breakdown of a camera-free eye-tracking system operating under two milliwatts?
- ▶ Why does calibration-free eye sensing tolerate glasses slippage, with no coordinate system and only eye speed?