Abstract
Angular velocity is a size-invariant motion metric that describes how fast an object is rotating. The computation of angular velocity relies on the scaling of local velocity information by the distance from the center of rotation. When we see a rotating object, what contributes to it's perceived rotation speed and do we see the objects angular velocity. To probe these questions we conducted a series of psychophysical experiments to determine how fast a line appears to rotate. The case of a rotating line is intriguing because the aperture-problem constrained measurement of local velocity is veridical—namely, as the line rotates, each point along the line's contour is moving orthogonally to the contour. We used the method of constant stimuli to determine the relative perceived speed of lines of different lengths and the method of adjustment to determine the perceived speed of a line that continuously changed length as it rotated. We found that a larger line appears to rotate faster than a shorter line. However, the difference we measured was less than would be expected if the perceived speed were based on the magnitudes of local component motion signals alone, suggesting a representation of the line's angular velocity. Additionally, we found a line that continuously changes size as it rotates appears to speed up as it gets larger and slow down as it gets smaller. This speed modulation is almost perfectly accounted for by the magnitudes of the local component motion signals, suggesting the absence of an angular velocity representation. This is likely to be the case because a computation of angular velocity requires the integration of a local motion signal with a computed distance from the center of rotation. This latter computation may be made difficult in the case of non-rigid motion as the distance is continually changing.
Meeting abstract presented at VSS 2018