3D Periodic Human Motion Reconstruction from 2D Motion Sequences

In this report, we present and evaluate a method of reconstructing three-dimensional (3D) periodic human motion from two-dimensional (2D) motion sequences. Based on a Fourier decomposition of a training set of 3D data, we construct a linear, morphable representation. Using this representation a lowdimensional linear model is learned by means of Principle Component Analysis (PCA). Twodimensional test data are now projected onto this model and the resulting 3D reconstructions are evaluated. We present two different simulations. In the first experiment, we assume the 2D projection matrix to be known. In the second experiment, the horizontal viewpoint is unknown and is being recovered from the data.

• Type: Journal
• Year: 2007
• Where: NECO
• Authors: Zonghua Zhang, Nikolaus F. Troje
• Keywords: Lowdimensional Linear Model, NECO 2007, Periodic Human Motion, Principle Component Analysis