Action Recognition using Motion Primitives

Publikation: Konferencebidrag uden forlag/tidsskriftPaper uden forlag/tidsskriftForskning

371 Downloads (Pure)

Resumé

The number of potential applications has made automatic recognition of human actions a very active research area. Different approaches have been followed based on trajectories through some state space. In this paper we also model an action as a trajectory through a state space, but we represent the actions as a sequence of temporal isolated instances, denoted primitives. These primitives are each defined by four features extracted from motion images. The primitives are recognized in each frame based on a trained classifier resulting in a sequence of primitives. From this sequence we recognize different temporal actions using a probabilistic Edit Distance method. The method is tested on different actions with and without noise and the results show recognizing rates of 88.7% and 85.5%, respectively.

OriginalsprogEngelsk
Publikationsdato2006
Antal sider10
StatusUdgivet - 2006
BegivenhedThe 15th Danish conference on pattern recognition and image analysis - Copenhagen, Danmark
Varighed: 24 aug. 200625 aug. 2006
Konferencens nummer: 15

Konference

KonferenceThe 15th Danish conference on pattern recognition and image analysis
Nummer15
LandDanmark
ByCopenhagen
Periode24/08/200625/08/2006

Fingerprint

Trajectories
Classifiers

Citer dette

Moeslund, T. B., Fihl, P., & Holte, M. B. (2006). Action Recognition using Motion Primitives. Afhandling præsenteret på The 15th Danish conference on pattern recognition and image analysis, Copenhagen, Danmark.
Moeslund, Thomas B. ; Fihl, Preben ; Holte, Michael Boelstoft. / Action Recognition using Motion Primitives. Afhandling præsenteret på The 15th Danish conference on pattern recognition and image analysis, Copenhagen, Danmark.10 s.
@conference{9e773f30a0ed11db8ed6000ea68e967b,
title = "Action Recognition using Motion Primitives",
abstract = "The number of potential applications has made automatic recognition of human actions a very active research area. Different approaches have been followed based on trajectories through some state space. In this paper we also model an action as a trajectory through a state space, but we represent the actions as a sequence of temporal isolated instances, denoted primitives. These primitives are each defined by four features extracted from motion images. The primitives are recognized in each frame based on a trained classifier resulting in a sequence of primitives. From this sequence we recognize different temporal actions using a probabilistic Edit Distance method. The method is tested on different actions with and without noise and the results show recognizing rates of 88.7{\%} and 85.5{\%}, respectively.",
keywords = "Human motion, Gesture recognition, Motion primitives",
author = "Moeslund, {Thomas B.} and Preben Fihl and Holte, {Michael Boelstoft}",
year = "2006",
language = "English",
note = "null ; Conference date: 24-08-2006 Through 25-08-2006",

}

Moeslund, TB, Fihl, P & Holte, MB 2006, 'Action Recognition using Motion Primitives' Paper fremlagt ved The 15th Danish conference on pattern recognition and image analysis, Copenhagen, Danmark, 24/08/2006 - 25/08/2006, .

Action Recognition using Motion Primitives. / Moeslund, Thomas B.; Fihl, Preben; Holte, Michael Boelstoft.

2006. Afhandling præsenteret på The 15th Danish conference on pattern recognition and image analysis, Copenhagen, Danmark.

Publikation: Konferencebidrag uden forlag/tidsskriftPaper uden forlag/tidsskriftForskning

TY - CONF

T1 - Action Recognition using Motion Primitives

AU - Moeslund, Thomas B.

AU - Fihl, Preben

AU - Holte, Michael Boelstoft

PY - 2006

Y1 - 2006

N2 - The number of potential applications has made automatic recognition of human actions a very active research area. Different approaches have been followed based on trajectories through some state space. In this paper we also model an action as a trajectory through a state space, but we represent the actions as a sequence of temporal isolated instances, denoted primitives. These primitives are each defined by four features extracted from motion images. The primitives are recognized in each frame based on a trained classifier resulting in a sequence of primitives. From this sequence we recognize different temporal actions using a probabilistic Edit Distance method. The method is tested on different actions with and without noise and the results show recognizing rates of 88.7% and 85.5%, respectively.

AB - The number of potential applications has made automatic recognition of human actions a very active research area. Different approaches have been followed based on trajectories through some state space. In this paper we also model an action as a trajectory through a state space, but we represent the actions as a sequence of temporal isolated instances, denoted primitives. These primitives are each defined by four features extracted from motion images. The primitives are recognized in each frame based on a trained classifier resulting in a sequence of primitives. From this sequence we recognize different temporal actions using a probabilistic Edit Distance method. The method is tested on different actions with and without noise and the results show recognizing rates of 88.7% and 85.5%, respectively.

KW - Human motion

KW - Gesture recognition

KW - Motion primitives

M3 - Paper without publisher/journal

ER -

Moeslund TB, Fihl P, Holte MB. Action Recognition using Motion Primitives. 2006. Afhandling præsenteret på The 15th Danish conference on pattern recognition and image analysis, Copenhagen, Danmark.