Abstract
Prior research has determined that our expectations of both target appearance and target location guide where we look when searching for an object in a scene. Expectations of location can be further divided according to the amount of visual information available to the observer. Probable locations could initially be predicted based only on the target identity but without any visual information from the scene ("Eyes closed"). Alternatively, expectations might only be generated after perceiving the scene layout or gist ("Eyes open"). The inter-observer consistency of these expectations was compared in three experiments by asking people to guess where a named object would be located by clicking within a picture frame. The results from Experiment 1 confirmed that people were consistent in their guesses when no visual information was given, but that responses became more tightly clustered when a brief preview of the scene was flashed beforehand, even when the target object was absent. In Experiment 2, the duration of the preview had a systematic effect on the consistency of observers’ guesses, with a longer preview making it more likely for participants to agree on an appropriate location. In Experiment 3, the eye fixations of new participants searching the same scenes were predictable based on the guesses from Experiments 1 and 2. Moreover, when participants were forced to make a speeded saccade they tended to fixate the location selected by guessers with less visual information. These results provide a simple method for quantifying expectations, their time-course, and their effects on visual exploration.
Meeting abstract presented at VSS 2013