Purchase this article with an account.
Tim Cornelissen, Kenneth Holmqvist, Melissa Vo; I couldn't help but notice: Irrelevant object-scene inconsistencies influence search for highly visible gabor patches. Journal of Vision 2017;17(10):565. doi: 10.1167/17.10.565.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The fact that we perform many searches through naturalistic scenes very efficiently on a daily basis suggests that object and scene identification (as well as their integration) require little attentional resources and might be obligatory in the sense that they are hard to suppress. Does obligatory scene processing also affect a purely perceptual task like gabor identification? Objects that do not fit the semantics of the scene (e.g. a toothbrush in an office) are typically fixated longer and more often than consistent controls. Here we overlaid a grid of gabor patches on a background scene that was irrelevant to the task of searching for a target gabor with a perfectly horizontal or vertical orientation. Some of the background images contained semantically inconsistent objects. To maximize their conspicuity gabors were outlined with a red box for easy saccadic targeting, thus avoiding the need to search through the background scene. Once fixated, figure-ground segmentation of the gabors on their gray background should be equally difficult for each element. For part of the irrelevant background scenes the target was overlaid on the critical object, whereas in other scenes the target was placed elsewhere and a distractor took its place on the object. Although time to first target fixation and RTs seemed unaffected by target placement or object consistency, analysis showed that participants looked longer at distractors placed on top of a semantically inconsistent object. Participants were also more likely to misjudge the target when a semantically inconsistent object was present in the scene. Replicating previous findings (Cornelissen & Vo, 2016) that scene and object identities are processed obligatorily and influence ongoing gaze behavior, we now extend these findings by ensuring high conspicuity and using only target present trials, while nevertheless showing that a purely perceptual task can be influenced by irrelevant background semantics.
Meeting abstract presented at VSS 2017
This PDF is available to Subscribers Only