Mobile Real-Time Eye-Tracking for Gaze-Aware Security Surveillance Support Systems

The validity of two commercial-off-the-shelf eye trackers, including a fixed system and a mobile device, was compared in the context of a gaze-based assistant tool implemented to aid security surveillance. Participants performed a security surveillance simulation that required them to detect incidents across multiple video feeds. Aid from the Scantracker system-pinpointing neglected cameras while detecting and correcting attentional tunneling and vigilance decrease-was provided using gaze measures. Despite the lower amount of valid data for the mobile system, both devices computed similar measures of surveillance performance. Findings speak to the feasibility of using mobile robust-to-movement decision-support systems based on real-time gaze data to help improve incident detection in security surveillance environments.