File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1109/TPAMI.2007.1167
- Scopus: eid_2-s2.0-37549055132
- PMID: 18084059
- WOS: WOS:000251580300007
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Gaussian process dynamical models for human motion
Title | Gaussian process dynamical models for human motion |
---|---|
Authors | |
Keywords | Animation Machine learning Motion Stochastic processes Time series analysis Tracking |
Issue Date | 2008 |
Citation | IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008, v. 30 n. 2, p. 283-298 How to Cite? |
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. © 2008 IEEE. |
Persistent Identifier | http://hdl.handle.net/10722/192717 |
ISSN | 2023 Impact Factor: 20.8 2023 SCImago Journal Rankings: 6.158 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Wang, JM | en_US |
dc.contributor.author | Fleet, DJ | en_US |
dc.contributor.author | Hertzmann, A | en_US |
dc.date.accessioned | 2013-11-20T04:58:59Z | - |
dc.date.available | 2013-11-20T04:58:59Z | - |
dc.date.issued | 2008 | en_US |
dc.identifier.citation | IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008, v. 30 n. 2, p. 283-298 | en_US |
dc.identifier.issn | 0162-8828 | en_US |
dc.identifier.uri | http://hdl.handle.net/10722/192717 | - |
dc.description.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. © 2008 IEEE. | en_US |
dc.language | eng | en_US |
dc.relation.ispartof | IEEE Transactions on Pattern Analysis and Machine Intelligence | en_US |
dc.subject | Animation | - |
dc.subject | Machine learning | - |
dc.subject | Motion | - |
dc.subject | Stochastic processes | - |
dc.subject | Time series analysis | - |
dc.subject | Tracking | - |
dc.title | Gaussian process dynamical models for human motion | en_US |
dc.type | Article | en_US |
dc.description.nature | link_to_subscribed_fulltext | en_US |
dc.identifier.doi | 10.1109/TPAMI.2007.1167 | en_US |
dc.identifier.pmid | 18084059 | - |
dc.identifier.scopus | eid_2-s2.0-37549055132 | en_US |
dc.identifier.volume | 30 | en_US |
dc.identifier.issue | 2 | en_US |
dc.identifier.spage | 283 | en_US |
dc.identifier.epage | 298 | en_US |
dc.identifier.isi | WOS:000251580300007 | - |
dc.identifier.issnl | 0162-8828 | - |