Using an avatar to develop a system for the predication of human body pose from moments

Abstract

Tracking people using movie sequences is not straightforward because of the human body's articulation and the complexity of a person's movements. In this paper we show how a person's 3D pose can be reconstructed by using corresponding silhouettes of video sequences from a monocular view. Currently, a virtual avatar is used to train the model for inferring the pose and a different avatar is used to produce novel examples not in the training set in order to evaluate the approach. The approach was subsequently tested using the silhouettes of walking people. © Springer-Verlag Berlin Heidelberg 2005.