Abstract
When an observer moves through the world a pattern of expanding motion known as optic flow is generated on the retina. Motion that is external to an observer perturbs the optic flow and can cause inaccuracies in heading estimation. Despite this, we are able to navigate through noisy environments, such as crowds. In crowds, external motion is produced by both the movement of other people though the scene, as well as their limb movements. One way that the visual system could simplify the computation of heading through crowds is by taking into account biological motion information, which provides cues as to the motion of the individuals within a crowd. To investigate this possibility, we measured heading accuracy during self-motion through crowds of point-light biological motion. We found that biological motion has a negative impact on heading estimation when people within a crowd move their limbs but do not move through space. When walkers moved independently through the scene, however, the presence of limb motion improved heading accuracy. This occurred for crowds containing both regular and perturbed walkers, suggesting that the observed effect does not stem from the processing of biological motion per se, but is likely produced by low-level cues inherent in the gait patterns of walkers.
Meeting abstract presented at VSS 2018