| Abstract |
|
In this paper we describe an approach for inferring the
body posture using a 3D visual-hull constructed from a
set of silhouettes. We introduce an appearance-based,
view-independent, 3D shape description for classifying
and identifying human posture using a support vector
machine. The proposed global shape description is
invariant to rotation, scale and translation and varies
continuously with 3D shape variations. This shape
representation is used for training a support vector
machine allowing the characterization of human body
postures from the computed visual hull. The main
advantage of the shape description is its ability to capture
human shape variation allowing the identification of body
postures across multiple people. The proposed method is
illustrated on a set of video streams of body postures
captured by four synchronous cameras.
|
 Additional Information
|
Citation:
Isaac Cohen, Hongxia Li,
"Inference of Human Postures by Classification of 3D Human Body Shape,"
amfg,
p. 74,
IEEE International Workshop on Analysis and Modeling of Faces and Gestures,
2003
|
