Abstract
When presenting human derived stimuli, it is often the case that motion and form are investigated separately. As a consequence, it has rarely been possible to systematically study the interaction of form and motion in person perception. We present a possible solution to this problem for investigating perception of biological motion. Modern computer animation techniques were used to generate hierarchical humanoid models that, driven by digitally captured human movement and in two experiments we explored the interaction of form and motion. Firstly we combined affective human movement with static facial expressions in order to assess the contribution of facial expression and motion to perception of angry, sad and happy emotion. The results of this experiment suggested that although faces dominated the perception of emotion under normal conditions, it is movement that modulated the intensity of emotion. However, in cases where facial expression presented a low salient signal, movement dominated perception in the angry condition. The second experiment was motivated by findings in biological motion that suggests perturbation of phasic relationships amongst limbs have consequences in the perception of form. We used human movement motion capture data to investigate the phase relationships in limbs — for instance the relationship between wrist, elbow and shoulder. We found that in general there is a short phase shift amongst limb parts and used this to perturb the human movement in both point-light displays and dynamic humanoid models. We will discuss the consequences of sensitivity to phase noise for point-light displays compared with the causally connected hierarchical model. The findings from both experiments show that for high-level perception, there are multifaceted perceptual consequences of combining form and motion, suggesting the different integration of the two types of information.