Purchase this article with an account.
Yoshiyuki Ueda; Visual ensemble perception is not invariant across object types. Journal of Vision 2017;17(10):1338. doi: 10.1167/17.10.1338.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We can summarize visually presented multiple objects at a glance and utilize it for further cognition (e.g., Ariely, 2001; Alvarez, 2011). This is called ensemble perception. In this study, to investigate invariance of ensemble perception, I asked participants to judge which of two categories in the majority of the group using four types of objects. If participants could perceive precious ensemble (e.g., mean or number) from multiple objects, the psychometric functions should be a sigmoid-shape regardless of object types. This is because mean of objects should be nearer to the template of one category than that of the other category when more than half of objects belong to this category, and participants can do the majority judgment correctly. Twelve objects belonged to either of two categories. The categories were decided based on color (red vs. green), shape (circle vs. square), size (larger vs. smaller than the exemplar), and facial expression (happy or angry vs. neutral) in each experiment. The identical objects were presented within each category in the experiments concerning to color and shape, whereas various objects were presented in those concerning to size and facial expressions. For color, shape, and size, participants could do the majority judgment accurately. Although the size judgment was less accurate than the color and shape judgments, the psychometric function of the size judgment showed a sigmoid-shape as the same as those of the color and shape judgments. However, the facial expression judgment was poor, and the psychometric function was linear. An additional experiment showed that participants did the majority judgment based on a part of faces rather than whole of them. These results suggest that ensemble perception is not invariant across object types and there would be the capacity limit for complicated objects.
Meeting abstract presented at VSS 2017
This PDF is available to Subscribers Only