Humans are Detected More Efficiently than Machines in the Context of Natural Scenes

Abstract

In the context of natural scenes, we recently showed that detecting humans among machine distractors is more efficient than detecting machines among human distractors (Mayer, Vuong, & Thornton, 2015). We concluded that the attentional system is tuned to efficiently process human form and motion. However, our results are also consistent with the possibility that discarding machine distractors is more efficient than discarding human distractors. In the present study, we replicated our previous visual search experiment but this time embedded targets amongst the same type of distractors; namely scenes displaying natural motion (e.g., billowing clouds, trees moving in the wind). Detecting humans among natural motion was more efficient than detecting machines among the same distractors as reflected in shallower search slopes, smaller intercepts, shorter first fixation durations on targets, and higher percentages of first fixations on targets. These findings are in line with efficient detection of human targets but not with efficient discarding of machine distractors.

Searching for illusory motion

“…To verify that this sample size would provide sufficient power to detect search slopes that deviated from zero -our main empirical question --we used the closest matching conditions from this previous work to conduct a priori power analyses. Observed effect sizes (Cohen's d) for the appropriate dynamic conditions were 2.67 (Mayer et al, 2015) and 1.11 (Mayer et al, 2017). Assuming required power of 0.8 and an alpha level of 0.05, these suggest minimum sample sizes of between 4 and 11 participants would be sufficient to detect RT x Set Size functions with nonzero slopes.…”

Comparable search efficiency for human and animal targets in the context of natural scenes

“…Using a standard visual search paradigm (Treisman & Gelade, 1980;Treisman & Souther, 1985;Wolfe, 1998;Wolfe & Horowitz, 2004, we showed that human targets embedded in natural scenes were located more efficiently than a range of complex, mechanical targets. This human search advantage, observed in terms of shallower search slopes, occurred in a standard search asymmetry design (in which humans and machines served as the targets and distractors for each other; Mayer et al, 2015), and also when such targets had to be found in the context of a third, common distractor class of moving natural objects, such as clouds and fire (Mayer et al, 2017). The shallower slopes in both our previous studies were complemented by higher proportions of first fixations landing on a human target and shorter ontarget fixation durations.…”

Editorial: Search: A New Perspective to Understand Cognitive Dynamics

“…Prominent candidates include animacy (Gao, Newman, & Scholl, ), threat (Ohman, Lundqvist, & Esteves, ), and faces and facial expressions (Frischen, Eastwood, & Smilek, ). In this issue, Mayer, Vuong, and Thornton () provide convincing evidence that a range of human motions (e.g., kicking a ball, doing the dishes) are easier to detect than machine motions (e.g., a moving carousel or an industrial sawing machine). By asking observers to search for human or machine motions against a background of “natural” distractor motions (e.g., moving clouds, flames), Mayer et al () are able to refute the possibility that the apparent advantage for human motions is the result of easier rejection of machine distractors (Wolfe & Horowitz, ).…”

“…The papers in this issue are quite diverse. They cover a variety of topics across the spectrum of contemporary search research: the prevalence effect (Horowitz, 2017); the importance of scene structure (Ueda, Kamakura, & Saiki, 2017); haptic search (Kaga, Kawaguchi, Mishina, Kita, & Watanabe, 2017); contextual cueing (Higuchi & Saiki, 2017;Makovski, 2017); foraging (Jóhannesson, Kristjánsson, & Thornton, 2017); inhibition of return (Niimi, Shimada, & Yokosawa, 2017); biological motion (Mayer, Vuong, & Thornton, 2017); and the cognitive effects of mobile phone use (Ito & Kawahara, 2017). They also cover a range of search behaviors, including tactile exploration (Kaga et al, 2017), vision-guided touch responses (Jóhannesson et al, 2017), oculomotor behavior (Ueda et al, 2017), in addition to standard button-press reaction time (RT) and accuracy measures.…”

“…In this issue, Mayer, Vuong, and Thornton (2017) provide convincing evidence that a range of human motions (e.g., kicking a ball, doing the dishes) are easier to detect than machine motions (e.g., a moving carousel or an industrial sawing machine). By asking observers to search for human or machine motions against a background of "natural" distractor motions (e.g., moving clouds, flames), Mayer et al (2017) are able to refute the possibility that the apparent advantage for human motions is the result of easier rejection of machine distractors (Wolfe & Horowitz, 2017). While human motions did not "pop out" in the traditional sense (i.e., flat RT × set size slope), the results do suggest an attentional advantage for characteristic human motions, possibly based on dynamic attention templates (sprites; Cavanagh, Labianca, & Thornton, 2001).…”