The camera system can see through the windshield of approaching vehicles in various time of day and weather conditions, and will generate appropriate security alerts.

Imagery was obtained by using multiple camera modalities to overcome various lighting, windshield, and weather combinations, supported by a novel data fusion system that used deep learning to associate different image types to the same person. The resulting fused data compared passengers to a photo gallery of pre-approved users, based on recent developments in facial recognition using recently developed Deep Polynomial Networks and related techniques.