Purchase this article with an account.
Phil McAleer, Helena M. Paterson, Barbara Mazzarino, Frank E. Pollick; Towards canonical views of animacy from scenes of human action. Journal of Vision 2005;5(8):936. doi: https://doi.org/10.1167/5.8.936.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
It is well known that social intention and meaning can be attributed to displays of moving geometric shapes, yet the cognitive processes that underlie this perception of animacy are, however, still open to debate. We have further explored this issue by making abstract displays of human movement using the multimedia analysis program Eyesweb (www.eyesweb.org). We first videotaped the movement of two actors from two viewing directions - an overhead and a side view. The actors performed everyday actions, interactions and scenes based on stimuli from the literature of animacy and causality research. These actions fell into the classes of highly scripted and loosely scripted actions. We then processed the video to obtain two visual conditions that depicted the motion centre of each actor as either a square remaining stable in size or changing in size according to the quantity of motion (QoM). The resultant stimuli depicted two white squares moving on a black background. We presented the four kinds of displays (side-view/overhead with no QoM/QoM) to naive observers with the task of rating the perception of animacy on a 9-point scale. The results suggest that there is a slight difference in viewpoint, with overhead being favoured and, that there is an effect of QoM in the side-view, yet this may be due to the change in size being interpreted as a depth cue. Finally, a trend was noted that displays with a greater control over the actors produced higher ratings of animacy than those displays obtained with a lower degree of control over the actors. Results are discussed in terms of variations in the creation of the stimuli and factors that influence the perception of animacy.
This PDF is available to Subscribers Only