Purchase this article with an account.
Benjamin Miller, George Andersen; Edge Detection through Kinetic Occlusion: The Role of Spatial and Temporal Integration. Journal of Vision 2017;17(10):1365. doi: 10.1167/17.10.1365.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Kinetic occlusion in visual scenes is a crucial source of information that allows us to successfully perceive objects and navigate the environment. In the current study we examined kinetic occlusion's role in the perception of curved edge boundaries. Previous work (Andersen and Cortese, 1989) had suggested that the spatial information provided by kinetic occlusion was a more critical factor in the perception of shapes than temporal information, so we sought to determine if the same was true for the discrimination of edges. Participants were presented pairs of displays with opaque objects occluding a moving background texture and asked to determine which of the two objects had a curved edge. Thresholds for curvature sensitivity were derived using a QUEST adaptive staircase procedure in a four (density of background texture) by four (velocity of background texture) design. Results indicated a main effect for both density of background texture elements and velocity of the moving texture. As either factor increased participants' sensitivity to curvature increased. A significant density by velocity interaction was also found, indicating a multiplicative effect of increasing one factor as the other increased. The effect of rate of accretion and deletion of texture was also separately analyzed. Surprisingly the analyses indicated that there were no significant differences between conditions that had the same rates of accretion and deletion. This would suggest that spatial information was not more important than temporal information. This finding suggests that the manner in which temporal and spatial information is integrated is different when trying to extract edge boundaries as compared to overall shape of an object.
Meeting abstract presented at VSS 2017
This PDF is available to Subscribers Only