Iris recognition is one of the most accurate biometric modalities, but its deployment has traditionally required high-resolution, close-range optics – keeping it out of the vehicle cabin. This talk presents how Smart Eye brought iris recognition into automotive environments by reducing the required iris resolution by nearly an order of magnitude, from 17 px/mm down to approximately 3 px/mm. Two innovations made this possible: temporal integration across video frames, which turns the low-resolution constraint into an advantage by leveraging temporal redundancy, and Smart Eye’s robust eye-tracking expertise, which provides the segmentation and localisation needed under challenging in-cabin conditions. Operating at occupant-camera resolutions also unlocks hardware-free multimodal fusion, combining face and iris into a system that outperforms either modality alone. The talk concludes by exploring implications for the in-cabin ecosystem: personalised settings, fleet sharing, and banking-grade authentication – all without adding a single sensor.
