Abstract
DeValois and DeValois (1991) have shown that motion within a stationary aperture leads to its perceptual displacement in the direction of motion. This distortion in localizing objects due to motion signals is not limited to the motion field itself, but occurs also for objects that are flashed in its vicinity (e.g. Whitney & Cavanagh, 2000). In earlier experiments we measured the size of this motion capture by comparing the perceived position of two targets flashed simultaneously on either side a moving Gabor (Friedrich et al., VSS '03), and by comparing the perceived position of a target flashed once during the motion and again after the motion has stopped (Friedrich & Mamassian, VSS '04). Problems with these methods were that when using two targets at different positions around the Gabor, we must know the size of the shift for one of the targets to use it as a reference. For the double-flashing of the same target, we have to take into account a possible motion aftereffect. To avoid these problems, we designed a Vernier alignment task. Three targets were vertically positioned and each one was placed next to a moving Gabor. The outer two Gabors were placed on one side of the targets (e.g. right) while the middle one was placed on the other side (left). In addition, the outer Gabors moved in a direction opposite to the middle one, so all three targets had the same position relative to the motion fields, either in front or behind. This configuration produced a perceptual misalignment of physically aligned targets. Targets were laterally shifted to find the shift leading to perfect perceptual alignment. For each observer, we took into account their baseline bias to localize the targets when the Gabors were stationary. For both targets in front and behind the motion fields, we found significant target shifts in the direction of motion. We extend this technique to measure motion capture for targets placed at various locations around a motion field.
We acknowledge the support of EC grant HPRN-CT-2002-00226.