Abstract
The “correspondence problem” refers to the task of maintaining object identity as the objects or the eyes move. Overall, the many studies using ambiguous motion displays (e.g., the motion quartet) find little or no influence of features (e.g., color or size) in resolving correspondence but a dominant influence of spatial and temporal proximity. In contrast, recent studies using the Ternus display show that feature information can play an important role in solving the correspondence problem (e.g., Hein & Moore, VSS 2009, VSS 2010). We investigated this conflict between the two paradigms by looking at the effect of distance on the strength of the feature influence in a variation of the Ternus display. Three equally spaced discs, aligned vertically, were presented in alternation with a second set displaced vertically so that the bottom two discs of the first set line up with the top two of the other set. The discs could be perceived as moving up and down together (group motion) or with one disc “jumping” from top end to the bottom while the other discs remained stationary (element motion). We varied the horizontal offset between the two sets of discs so that they would move slightly sideways as well as up and down. We biased the percept toward element motion by matching features of surface polarity, size or spatial frequency. We found that the percept of element motion decreased as the horizontal shift increased and that the distance at which the decrease started differed for the different features. These findings suggest that feature information contributes to correspondence but only over a small spatial range. This limited range explains why the Ternus display shows a feature effect—the central discs are typically spatially superimposed—whereas other apparent motion displays use large spatial offsets and show little or no feature effects.
Supported by Chaire d'Excellence ANR grant to PC.