Locally time-invariant models of human activities using trajectories on the grassmannian

Pavan Turaga, Rama Chellappa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

31 Citations (Scopus)

Abstract

Human activity analysis is an important problem in computer vision with applications in surveillance and summarization and indexing of consumer content. Complex human activities are characterized by non-linear dynamics that make learning, inference and recognition hard. In this paper, we consider the problem of modeling and recognizing complex activities which exhibit time-varying dynamics. To this end, we describe activities as outputs of linear dynamic systems (LDS) whose parameters vary with time, or a Time-Varying Linear Dynamic System (TV-LDS). We discuss parameter estimation methods for this class of models by assuming that the parameters are locally time-invariant. Then, we represent the space of LDS models as a Grassmann manifold. Then, the TV-LDS model is defined as a trajectory on the Grassmann manifold. We show how trajectories on the Grassmannian can be characterized using appropriate distance metrics and statistical methods that reflect the underlying geometry of the manifold. This results in more expressive and powerful models for complex human activities. We demonstrate the strength of the framework for activity-based summarization of long videos and recognition of complex human actions on two datasets.

Original languageEnglish (US)
Title of host publication2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009
Pages2435-2441
Number of pages7
DOIs
StatePublished - 2009
Externally publishedYes
Event2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009 - Miami, FL, United States
Duration: Jun 20 2009Jun 25 2009

Other

Other2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009
CountryUnited States
CityMiami, FL
Period6/20/096/25/09

Fingerprint

Dynamical systems
Trajectories
Parameter estimation
Computer vision
Statistical methods
Geometry

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Biomedical Engineering

Cite this

Turaga, P., & Chellappa, R. (2009). Locally time-invariant models of human activities using trajectories on the grassmannian. In 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009 (pp. 2435-2441). [5206710] https://doi.org/10.1109/CVPRW.2009.5206710

Locally time-invariant models of human activities using trajectories on the grassmannian. / Turaga, Pavan; Chellappa, Rama.

2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009. 2009. p. 2435-2441 5206710.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Turaga, P & Chellappa, R 2009, Locally time-invariant models of human activities using trajectories on the grassmannian. in 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009., 5206710, pp. 2435-2441, 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009, Miami, FL, United States, 6/20/09. https://doi.org/10.1109/CVPRW.2009.5206710
Turaga P, Chellappa R. Locally time-invariant models of human activities using trajectories on the grassmannian. In 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009. 2009. p. 2435-2441. 5206710 https://doi.org/10.1109/CVPRW.2009.5206710
Turaga, Pavan ; Chellappa, Rama. / Locally time-invariant models of human activities using trajectories on the grassmannian. 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009. 2009. pp. 2435-2441
@inproceedings{57f6b6fb42564cf1a5465000bddef482,
title = "Locally time-invariant models of human activities using trajectories on the grassmannian",
abstract = "Human activity analysis is an important problem in computer vision with applications in surveillance and summarization and indexing of consumer content. Complex human activities are characterized by non-linear dynamics that make learning, inference and recognition hard. In this paper, we consider the problem of modeling and recognizing complex activities which exhibit time-varying dynamics. To this end, we describe activities as outputs of linear dynamic systems (LDS) whose parameters vary with time, or a Time-Varying Linear Dynamic System (TV-LDS). We discuss parameter estimation methods for this class of models by assuming that the parameters are locally time-invariant. Then, we represent the space of LDS models as a Grassmann manifold. Then, the TV-LDS model is defined as a trajectory on the Grassmann manifold. We show how trajectories on the Grassmannian can be characterized using appropriate distance metrics and statistical methods that reflect the underlying geometry of the manifold. This results in more expressive and powerful models for complex human activities. We demonstrate the strength of the framework for activity-based summarization of long videos and recognition of complex human actions on two datasets.",
author = "Pavan Turaga and Rama Chellappa",
year = "2009",
doi = "10.1109/CVPRW.2009.5206710",
language = "English (US)",
isbn = "9781424439935",
pages = "2435--2441",
booktitle = "2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009",

}

TY - GEN

T1 - Locally time-invariant models of human activities using trajectories on the grassmannian

AU - Turaga, Pavan

AU - Chellappa, Rama

PY - 2009

Y1 - 2009

N2 - Human activity analysis is an important problem in computer vision with applications in surveillance and summarization and indexing of consumer content. Complex human activities are characterized by non-linear dynamics that make learning, inference and recognition hard. In this paper, we consider the problem of modeling and recognizing complex activities which exhibit time-varying dynamics. To this end, we describe activities as outputs of linear dynamic systems (LDS) whose parameters vary with time, or a Time-Varying Linear Dynamic System (TV-LDS). We discuss parameter estimation methods for this class of models by assuming that the parameters are locally time-invariant. Then, we represent the space of LDS models as a Grassmann manifold. Then, the TV-LDS model is defined as a trajectory on the Grassmann manifold. We show how trajectories on the Grassmannian can be characterized using appropriate distance metrics and statistical methods that reflect the underlying geometry of the manifold. This results in more expressive and powerful models for complex human activities. We demonstrate the strength of the framework for activity-based summarization of long videos and recognition of complex human actions on two datasets.

AB - Human activity analysis is an important problem in computer vision with applications in surveillance and summarization and indexing of consumer content. Complex human activities are characterized by non-linear dynamics that make learning, inference and recognition hard. In this paper, we consider the problem of modeling and recognizing complex activities which exhibit time-varying dynamics. To this end, we describe activities as outputs of linear dynamic systems (LDS) whose parameters vary with time, or a Time-Varying Linear Dynamic System (TV-LDS). We discuss parameter estimation methods for this class of models by assuming that the parameters are locally time-invariant. Then, we represent the space of LDS models as a Grassmann manifold. Then, the TV-LDS model is defined as a trajectory on the Grassmann manifold. We show how trajectories on the Grassmannian can be characterized using appropriate distance metrics and statistical methods that reflect the underlying geometry of the manifold. This results in more expressive and powerful models for complex human activities. We demonstrate the strength of the framework for activity-based summarization of long videos and recognition of complex human actions on two datasets.

UR - http://www.scopus.com/inward/record.url?scp=70450207848&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70450207848&partnerID=8YFLogxK

U2 - 10.1109/CVPRW.2009.5206710

DO - 10.1109/CVPRW.2009.5206710

M3 - Conference contribution

AN - SCOPUS:70450207848

SN - 9781424439935

SP - 2435

EP - 2441

BT - 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2009

ER -