Abstract
Previous experiments using degraded target objects have shown that recognition is facilitated by contextual information (Bar & Aminoff, 1996; Barenholtz, In Press). Yet, it is not known how effective this contextual information is relative to other sources of information, such as an object's intrinsic visual features. To address this, we performed four experiments using rendered scenes with novel objects, in which the features and locations of the objects could be experimentally manipulated. In all four experiments, participants first performed a visual search task, searching for uniquely shaped target objects within the scene. The objects' colors and location within the context were statistically manipulated during the search phase. We then tested participants' tendency to use their knowledge of the location/color information for the purposes of recognition when the object's image was degraded due to blurring, eliminating the shape information. In Experiment 1, we found that, in the absence of any diagnostic intrinsic features, participants identified objects based purely on their locations within the scene. In Experiment 2, we found that participants combined an intrinsic feature, color, with contextual location in order to uniquely specify an object. In Experiment 3, we found that when an object's color and location information were in conflict, participants identified the object using both sources of information equally. Finally, Experiment 4 found that participants used whichever source of information was more statistically reliable—either color or location— in order to identify the target object; however reliable location was given higher priority than reliable color. Overall, these experiments show that the context of objects may be a potent source for object identification, playing as important a role as intrinsic object features under some conditions.
Meeting abstract presented at VSS 2014