Perception of biological motion by intermediate-level motion features
In this thesis, it is proposed that the visual perception of biological motion, both in natural images and point-light displays (PLD), depends upon the detection of certain intermediate level motion features . These motion features characterize the spatiotemporal patterns that describe the relative motion of body parts. Detection of these features may be orientation sensitive - therefore accounting for increased sensitivity to biological motion in upright displays. We measured human observers' sensitivity to perturbations of these intermediate level spatiotemporal features in point-light displays of upright and inverted walkers. We found that observers are more sensitive to perturbation of certain features in upright displays but more sensitive to other features in inverted displays. The features with greater sensitivity in upright displays describe the relative motion of adjacent limb segments (e.g., left thigh/right thigh, and thigh/leg). In contrast, perturbations to a feature that describes the angular velocity of the limbs result in greater sensitivity in inverted displays. We hypothesize that this procedure allows identification of intermediate level features used by the visual system for biological motion recognition. ^ These spatiotemporal motion features closely correspond to the first few principal components (PC) of the kinematic data describing gait. We describe a methodology for classification of gait (walk, run, jog, etc.) and recognition of individuals based on gait using two successive stages of principal component analysis (PCA) on kinematic data. The first few PCs provide a representation of an individual gait as trajectory along a low-dimensional manifold in PC space. A second stage of PCA captures variability in the shape of this manifold across individuals or gaits. This simple eigenspace based analysis is capable of accurate classification across subjects. ^
Biology, Neuroscience|Engineering, Biomedical
"Perception of biological motion by intermediate-level motion features"
(January 1, 2006).
Dissertations available from ProQuest.