Visual perception of biological motion by form: a template-matching analysis.

Biological motion perception is referred to as the ability to recognize a moving human figure from no more than a few moving point lights. Such point-light stimuli contain limited form information about the shape of the body and local image motion signals from the moving points. The contributions of form and motion to the vivid perception of point-light displays are subject to controversy in the discussion. While some studies claim that local motion signals are critical, others emphasize the role of global form cues. Here, we present a template-matching approach to investigate the role of global form analysis. We used a template-matching method that derives biological motion exclusively from form information. The algorithm used static postures monitored from walking humans as stored templates. We compared the simulation results to psychophysical experiments with the commonly used point-light walker and a variant point-light walker with near-absent local motion signals. The common result in all experiments was a high correlation between simulation results and psychophysical data. The results show that the limited form information in point-light stimuli might be sufficient to perceive biological motion. We suggest that it is possible for humans to extract the sparse form information in point-light walkers and to use it to perceive biological motion by integrating dynamic form information over time.