Adaptation of a Gaze-Aware Security Surveillance Support Tool for Augmented Reality

Operators monitoring closed-circuit television systems face many error-prone situations given the cognitive challenges they experience. The Scantracker is a prototype tool developed to support surveillance, relying on live oculometric data to trigger notifications that mitigate camera negligence, attention tunneling and vigilance decrement. To be robust and flexible, this tool must however collect valid gaze measures at a sufficient sampling rate and allow the display of surveillance-supportive visual notifications. This paper documents the integration of the Scantracker with the Microsoft HoloLens 2 augmented reality (AR) system embarked with eye tracking and reports a validity study performed in a high-fidelity surveillance simulation. The AR solution led to superior gaze validity than previous mobile or fixed oculometer integration. Implications of the solution integration are discussed and compared with the other implementations.