Gaussian Process Dynamical Models for Human Motion Export

@article{wang2008, abstract = {We introduce Gaussian process dynamical models (GPDM) for nonlinear time series analysis, with applications to learning models of human pose and motion from high-dimensionalmotion capture data. A GPDM is a latent variable model. It comprises a low-dimensional latent space with associated dynamics, and a map from the latent space to an observation space. We marginalize out the model parameters in closed-form, using Gaussian process priors for both the dynamics and the observation mappings. This results in a non-parametric model for dynamical systems that accounts for uncertainty in the model. We demonstrate the approach, and compare four learning algorithms on human motion capture data in which each pose is 50-dimensional. Despite the use of small data sets, the GPDM learns an effective representation of the nonlinear dynamics in these spaces.}, address = {Washington, DC, USA}, author = {Wang, Jack M. and Fleet, David J. and Hertzmann, Aaron}, booktitle = {Pattern Analysis and Machine Intelligence, IEEE Transactions on}, citeulike-article-id = {3504557}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1340231}, citeulike-linkout-1 = {http://dx.doi.org/10.1109/TPAMI.2007.1167}, citeulike-linkout-2 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4359316}, doi = {10.1109/TPAMI.2007.1167}, issn = {0162-8828}, journal = {IEEE Trans. Pattern Anal. Mach. Intell.}, keywords = {dim\_reduction, gplvm}, number = {2}, pages = {283--298}, posted-at = {2009-09-29 13:50:00}, priority = {5}, publisher = {IEEE Computer Society}, title = {Gaussian Process Dynamical Models for Human Motion}, url = {http://dx.doi.org/10.1109/TPAMI.2007.1167}, volume = {30}, year = {2008} }

TY - JOUR ID - wang2008 L3 - citeulike-article-id:3504557 N2 - We introduce Gaussian process dynamical models (GPDM) for nonlinear time series analysis, with applications to learning models of human pose and motion from high-dimensionalmotion capture data. A GPDM is a latent variable model. It comprises a low-dimensional latent space with associated dynamics, and a map from the latent space to an observation space. We marginalize out the model parameters in closed-form, using Gaussian process priors for both the dynamics and the observation mappings. This results in a non-parametric model for dynamical systems that accounts for uncertainty in the model. We demonstrate the approach, and compare four learning algorithms on human motion capture data in which each pose is 50-dimensional. Despite the use of small data sets, the GPDM learns an effective representation of the nonlinear dynamics in these spaces. IS - 2 JF - IEEE Trans. Pattern Anal. Mach. Intell. SN - 0162-8828 AD - Washington, DC, USA EP - 298 TI - Gaussian Process Dynamical Models for Human Motion PB - IEEE Computer Society VL - 30 T2 - Pattern Analysis and Machine Intelligence, IEEE Transactions on SP - 283 KW - dim_reduction KW - gplvm AU - Wang, Jack AU - Fleet, David AU - Hertzmann, Aaron PY - 2008/// UR - http://dx.doi.org/10.1109/TPAMI.2007.1167 ER -