Abstract
Recognition of human action from point-light displays has occupied a unique position in the study of visual perception and cognition. The attraction of these stimuli is their facility to omit irrelevant aspects of the human form that might make recognition or detection trivial. In some situations, however, point-light stimuli are too abstract to reliably depict person characteristics and there is an increasing trend towards using animated models that incorporate additional form information. We performed two experiments to compare the perception of point-light (PL) displays to solid body model (SBM) displays that depict the same captured human movements. Solid body models are well-developed animation tools that provide precise control over the presentation of form as a volumetric solid. In comparing PLs to SBMs we expect that performance in judgements about PLs would differ from those about SBMs, in cases where shape informs the required judgement. In Experiment 1 observers were asked to categorise the affect of actors performing knocking movements with 10 emotions. Results showed little difference in proportion correct responses or in a psychological space of affect, based on multi-dimensional scaling of the confusion matrix of responses. This indicated that body shape does not inform perception of emotion. SBMs had, however, been presented with neutral facial expressions and we might expect that varying facial expression would strongly effect the interpretation of motion cues to emotion. In experiment 2 observers were asked to make judgements about the weight of an invisible, pushed object. Results showed that a greater range of weights was used to describe objects pushed by SBMs than the same objects pushed in PL displays. This suggests form information, when combined with motion information, modulates the perception of perceived weight to a greater extent than motion alone.