Training and transfer of training in the search for camouflaged targets

Abstract

950How do humans go about finding a search target when, through design (e.g., a military tank in the desert) or evolution (e.g., a lion in a savanna), it blends in with the background? Visual search has been extensively studied and a great deal has been learned about the guidance of attention during search, including how attention is controlled by bottom-up factors (e.g., Proulx, 2007;Theeuwes, 1992Theeuwes, , 2004Yantis & Jonides, 1984), top-down guidance (e.g., Bacon & Egeth, 1994;Chen & Zelinsky, 2006;Folk, Remington, & Johnston, 1992), and various memory mechanisms (e.g., Boot, McCarley, Kramer, & Peterson, 2004;Brockmole & Henderson, 2006;Chun & Jiang, 1998;Klein & MacInnes, 1999;Neider & Zelinsky, 2006b). Much of what we know about search has been derived from experiments in which target and distractors are easy to segment from the background (see Wolfe, 1994, for a review). However, many real-world search situations contain items that need to be segmented from the background to some degree to allow for efficient visual search. This is especially true in the search for camouflaged targets.As outlined by Wolfe, Oliva, Horowitz, Butcher, and Bompas (2002), the visual system faces a number of potentially important challenges when target and distractor objects are placed within the context of a visually complex background. These problems include more difficult preattentive segmentation of objects from the background and the creation of "illusory objects" by noise in the background. Additionally, if segmentation is imperfect, attentional selection may be more difficult and time consuming. Imperfect segmentation might also result in noisy object representations being passed along to higher visual stages, such as those responsible for object recognition. Through a series of cleverly designed search experiments, Wolfe and colleagues found that the main cost of background complexity was associated with delayed information accrual affecting later target recognition stages. However, in extreme cases of item-background similarity, item-by-item selection was slowed. Importantly to the present context, background complexity had a negative impact on search performance in all experiments, highlighting the difficulties associated with complex searches in which the search items are not easily distinguished from the background.To follow up on the findings of Wolfe et al. (2002), Neider and Zelinsky (2006a) directly examined the effect of target-background similarity. Neider and Zelinsky (2006b) had participants search through displays containing many children's toys and asked them to indicate whether or not a specific toy was present. Camouflage displays were created by taking a square-shaped portion of the target image and repeating it again and again to create a background. Search items were then superimposed on this background (Figure 1). This created a search task in which the features How do observers become proficient at finding camouflaged targets? Does gaining proficiency in locating a target in one camouflage search...

Training and Transfer of Training in Rapid Visual Search for Camouflaged Targets

“…Following this initial evaluation, all participants completed five training sessions with 360 trials per session for a total of 1,800 training trials. As in previous studies [1], [4], half of the participants were trained with a non-camouflage background while the other half trained with a camouflage background. Following training, all participants again performed the same two search tasks as in the pre-training test session.…”

“…To fill this gap in the literature, Boot, Neider and Kramer explored whether training could improve camouflage target sensitivity and whether training participants to detect one set of camouflaged targets could engender transfer to a novel set of camouflaged targets ( [1]; also see [4], for a study of age-related differences in training and transfer of search for camouflaged targets). Specifically, they had one group of participants train on a camouflage visual search task (using a nearly identical paradigm to Neider and Zelinsky [3], and another group of participants train on a visual search task using the same objects, but arrayed on a homogenous background (no camouflage).…”

“…To do so, we trained participants to search for camouflaged or non-camouflaged targets, but whereas previous studies used self-terminating search tasks [1], [4], we used a rapid presentation paradigm where participants only viewed the display for 150 ms before making a target presence response. Participants were also required to maintain central fixation at all times and this was ensured using an eye tracker.…”

“…Previous examinations of search under camouflage conditions have reported that performance improves with training and that training can engender near perfect transfer to similar, but novel camouflage-type displays [1]. What remains unclear, however, are the cognitive mechanisms underlying these training improvements and transfer benefits.…”

Improving the Efficacy of Security Screening Tasks: A Review of Visual Search Challenges and Ways to Mitigate Their Adverse Effects

“…For example, searchers are trained to identify objects from a wide variety of noncanonical viewpoints (e.g., Koller et al, 2008), such as a knife rotated with the base aimed at the screener. Effective training regimens for identifying prohibited items and breaking camouflage remain a continuing area of research (e.g., Boot, Neider, & Kramer, 2009;Chen & Hegdé, 2012), and these efforts should help improve future performance. In addition, implementing a few specific procedural efforts may be highly effective.…”