Using Eye Movements to Understand how Security Screeners Search for Threats in X-Ray Baggage

Abstract

There has been an increasing drive to understand failures in searches for weapons and explosives in X-ray baggage screening. Tracking eye movements during the search has produced new insights into the guidance of attention during the search, and the identification of targets once they are fixated. Here, we review the eye-movement literature that has emerged on this front over the last fifteen years, including a discussion of the problems that real-world searchers face when trying to detect targets that could do serious harm to people and infrastructure.

Improved X-ray baggage screening sensitivity with ‘targetless’ search training

“…Experiment 3 also introduced a key element of real-world baggage search that was missing from Experiments 1 and 2, which is a greater degree of spatial overlap within our testing stimuli. In the variant of the baggage search task used here, a proportion of the testing stimuli incorporated a greater degree of spatial overlap in to better simulate the real-world baggage search environment (Donnelly et al, 2019 ).…”

“…Previous studies of distractor templates have involved basic laboratory search tasks, but it is difficult to predict how effective they may be in baggage search due to the unique combination of challenges to the human observer posed by this task. Perhaps the most fundamental of these is that baggage X-rays lack a regular structure (Donnelly et al, 2019 ). In laboratory tasks, everyday searches and medical imaging, the search environment involves some predictable structure, e.g.…”

“…Baggage X-rays typically follow a standard colour mapping, whereby objects are coloured based upon their density and absorption of X-ray radiation (Donnelly et al, 2019 ). By this mapping, metallic objects (higher density) are coloured blue and organic objects (lower density) are orange.…”

“…While the effect of low prevalence in real-world baggage screening may be ameliorated through procedures that artificially increase threat prevalence (e.g. Threat Image Projection), it nonetheless remains a challenge for screeners (Donnelly et al, 2019 ). A related literature exists on satisfaction of search effects (also referred to as subsequent search misses; Adamo et al, 2018 ; Cain et al, 2013 ; Fleck et al, 2010 ), whereby searchers are less likely to detect subsequent targets after finding an initial target.…”

“…Threat Image Projection), it nonetheless remains a challenge Muhl-Richardson et al Cogn. Research (2021) 6:33 for screeners (Donnelly et al, 2019). A related literature exists on satisfaction of search effects (also referred to as subsequent search misses; Adamo et al, 2018;Cain et al, 2013;Fleck et al, 2010), whereby searchers are less likely to detect subsequent targets after finding an initial target.…”