Abstract
Many studies have shown that position uncertainty impairs performance while detecting and discriminating a stimulus. However, few of these studies have specifically examined intrinsic position uncertainty (IPU), which cannot be directly manipulated. Instead, most investigations have focused on extrinsic position uncertainty (EPU), which can be manipulated. In a recent study, we developed an experimental technique that modulates the distribution of clutter in synthetic noise displays while keeping EPU fixed. Because clutter at irrelevant locations only degrades search performance via IPU, we were able to show how IPU limits performance for overt search in synthetic displays. In the current study, we investigated whether IPU similarly limits overt search performance in natural images. Observers completed a search task that required locating an object (e.g., a key) in a natural image representing the contents of a bag. At the start of each trial, observers saw cues representing potential target locations. After completing their search, they responded by reporting either the perceived location of the target (if it was present) or by reporting its absence. However, rather than imposing synthetic clutter, as in the previous task, we used an existing measure (Bravo & Farid, 2008) to quantify the clutter for a set of natural images. We then modulated the clutter distribution by assigning high clutter images to one condition and low clutter images to another. We independently varied the number of potential target locations to manipulate extrinsic position uncertainty (EPU). Finally, we manipulated target presence across images to examine its interaction with the amount of clutter. Our results show that the amount of clutter in natural images degrades search performance irrespective of EPU, suggesting that IPU limits performance for overt search in natural images.
Meeting abstract presented at VSS 2017