Recognition of Action as a Bayesian Parameter Estimation Problem over Time

Research output: Contribution to book/anthology/report/conference proceedingBook chapterResearchpeer-review

Abstract

In this paper we will discuss two problems related to action recognition: The first problem is the one of identifying in a
surveillance scenario whether a person is walking or running and in what rough direction. The second problem is concerned with the recovery of action primitives from observed complex actions. Both problems will be discussed within a statistical framework. Bayesian propagation over time offers a framework to treat likelihood observations at each time step and the dynamics between the time steps in a unified manner. The first problem will be approached as a patter recognition and tracking task by a Bayesian propagation of the likelihoods. The latter problem will be 
approached by explicitly specifying the dynamics while the likelihood measure will give a measure how good each dynamical model fit at each time step. Extensive experimental results show the applicability of the Bayesian framework for action recognition and round up our discussion.
Original languageEnglish
Title of host publicationHuman Motion : Understanding, Modelling, Capture and Animation
EditorsB Rosenhahn, R Klette, D Metaxas
Number of pages25
Place of PublicationHeidelberg
PublisherIEEE Computer Society Press
Publication date2007
ISBN (Print)978-1-4020-6692-4
Publication statusPublished - 2007
SeriesComputational Imaging and Vision
Number36

Fingerprint

Parameter estimation
Recovery

Cite this

Krüger, V. (2007). Recognition of Action as a Bayesian Parameter Estimation Problem over Time. In B. Rosenhahn, R. Klette, & D. Metaxas (Eds.), Human Motion: Understanding, Modelling, Capture and Animation Heidelberg: IEEE Computer Society Press. Computational Imaging and Vision, No. 36
Krüger, Volker. / Recognition of Action as a Bayesian Parameter Estimation Problem over Time. Human Motion: Understanding, Modelling, Capture and Animation. editor / B Rosenhahn ; R Klette ; D Metaxas. Heidelberg : IEEE Computer Society Press, 2007. (Computational Imaging and Vision; No. 36).
@inbook{e42e97e09fd911db8ed6000ea68e967b,
title = "Recognition of Action as a Bayesian Parameter Estimation Problem over Time",
abstract = "In this paper we will discuss two problems related to action recognition: The first problem is the one of identifying in asurveillance scenario whether a person is walking or running and in what rough direction. The second problem is concerned with the recovery of action primitives from observed complex actions. Both problems will be discussed within a statistical framework. Bayesian propagation over time offers a framework to treat likelihood observations at each time step and the dynamics between the time steps in a unified manner. The first problem will be approached as a patter recognition and tracking task by a Bayesian propagation of the likelihoods. The latter problem will be approached by explicitly specifying the dynamics while the likelihood measure will give a measure how good each dynamical model fit at each time step. Extensive experimental results show the applicability of the Bayesian framework for action recognition and round up our discussion.",
author = "Volker Kr{\"u}ger",
year = "2007",
language = "English",
isbn = "978-1-4020-6692-4",
editor = "B Rosenhahn and R Klette and D Metaxas",
booktitle = "Human Motion",
publisher = "IEEE Computer Society Press",
address = "United States",

}

Krüger, V 2007, Recognition of Action as a Bayesian Parameter Estimation Problem over Time. in B Rosenhahn, R Klette & D Metaxas (eds), Human Motion: Understanding, Modelling, Capture and Animation. IEEE Computer Society Press, Heidelberg, Computational Imaging and Vision, no. 36.

Recognition of Action as a Bayesian Parameter Estimation Problem over Time. / Krüger, Volker.

Human Motion: Understanding, Modelling, Capture and Animation. ed. / B Rosenhahn; R Klette; D Metaxas. Heidelberg : IEEE Computer Society Press, 2007.

Research output: Contribution to book/anthology/report/conference proceedingBook chapterResearchpeer-review

TY - CHAP

T1 - Recognition of Action as a Bayesian Parameter Estimation Problem over Time

AU - Krüger, Volker

PY - 2007

Y1 - 2007

N2 - In this paper we will discuss two problems related to action recognition: The first problem is the one of identifying in asurveillance scenario whether a person is walking or running and in what rough direction. The second problem is concerned with the recovery of action primitives from observed complex actions. Both problems will be discussed within a statistical framework. Bayesian propagation over time offers a framework to treat likelihood observations at each time step and the dynamics between the time steps in a unified manner. The first problem will be approached as a patter recognition and tracking task by a Bayesian propagation of the likelihoods. The latter problem will be approached by explicitly specifying the dynamics while the likelihood measure will give a measure how good each dynamical model fit at each time step. Extensive experimental results show the applicability of the Bayesian framework for action recognition and round up our discussion.

AB - In this paper we will discuss two problems related to action recognition: The first problem is the one of identifying in asurveillance scenario whether a person is walking or running and in what rough direction. The second problem is concerned with the recovery of action primitives from observed complex actions. Both problems will be discussed within a statistical framework. Bayesian propagation over time offers a framework to treat likelihood observations at each time step and the dynamics between the time steps in a unified manner. The first problem will be approached as a patter recognition and tracking task by a Bayesian propagation of the likelihoods. The latter problem will be approached by explicitly specifying the dynamics while the likelihood measure will give a measure how good each dynamical model fit at each time step. Extensive experimental results show the applicability of the Bayesian framework for action recognition and round up our discussion.

M3 - Book chapter

SN - 978-1-4020-6692-4

BT - Human Motion

A2 - Rosenhahn, B

A2 - Klette, R

A2 - Metaxas, D

PB - IEEE Computer Society Press

CY - Heidelberg

ER -

Krüger V. Recognition of Action as a Bayesian Parameter Estimation Problem over Time. In Rosenhahn B, Klette R, Metaxas D, editors, Human Motion: Understanding, Modelling, Capture and Animation. Heidelberg: IEEE Computer Society Press. 2007. (Computational Imaging and Vision; No. 36).