The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Cet article aborde le problème de la reconnaissance d'actions invariantes en utilisant des trajectoires 2D de points de repère sur le corps humain. Il s'agit d'une tâche difficile car pour une catégorie d'action spécifique, les observations 2D de différentes instances peuvent être extrêmement différentes en raison des différences de point de vue et des changements de vitesse. En supposant que l'exécution d'une action peut être approchée par une combinaison linéaire dynamique d'un ensemble de formes de base, une nouvelle méthode de reconnaissance d'action humaine invariante de vue est proposée, basée sur la factorisation matricielle non rigide et les modèles de Markov cachés (HMM). Nous montrons que les faibles coefficients de poids dimensionnel des formes de base par factorisation non rigide de matrice de mesure contiennent les informations clés pour la reconnaissance des actions, quel que soit le changement de point de vue. Sur la base des caractéristiques discriminantes extraites, le HMM est utilisé pour la modélisation dynamique temporelle et la classification robuste des actions. La méthode proposée est testée en utilisant des séquences réelles et des performances prometteuses sont obtenues.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Xi LI, Kazuhiro FUKUI, "View Invariant Human Action Recognition Based on Factorization and HMMs" in IEICE TRANSACTIONS on Information,
vol. E91-D, no. 7, pp. 1848-1854, July 2008, doi: 10.1093/ietisy/e91-d.7.1848.
Abstract: This paper addresses the problem of view invariant action recognition using 2D trajectories of landmark points on human body. It is a challenging task since for a specific action category, the 2D observations of different instances might be extremely different due to varying viewpoint and changes in speed. By assuming that the execution of an action can be approximated by dynamic linear combination of a set of basis shapes, a novel view invariant human action recognition method is proposed based on non-rigid matrix factorization and Hidden Markov Models (HMMs). We show that the low dimensional weight coefficients of basis shapes by measurement matrix non-rigid factorization contain the key information for action recognition regardless of the viewpoint changing. Based on the extracted discriminative features, the HMMs is used for temporal dynamic modeling and robust action classification. The proposed method is tested using real life sequences and promising performance is achieved.
URL: https://global.ieice.org/en_transactions/information/10.1093/ietisy/e91-d.7.1848/_p
Copier
@ARTICLE{e91-d_7_1848,
author={Xi LI, Kazuhiro FUKUI, },
journal={IEICE TRANSACTIONS on Information},
title={View Invariant Human Action Recognition Based on Factorization and HMMs},
year={2008},
volume={E91-D},
number={7},
pages={1848-1854},
abstract={This paper addresses the problem of view invariant action recognition using 2D trajectories of landmark points on human body. It is a challenging task since for a specific action category, the 2D observations of different instances might be extremely different due to varying viewpoint and changes in speed. By assuming that the execution of an action can be approximated by dynamic linear combination of a set of basis shapes, a novel view invariant human action recognition method is proposed based on non-rigid matrix factorization and Hidden Markov Models (HMMs). We show that the low dimensional weight coefficients of basis shapes by measurement matrix non-rigid factorization contain the key information for action recognition regardless of the viewpoint changing. Based on the extracted discriminative features, the HMMs is used for temporal dynamic modeling and robust action classification. The proposed method is tested using real life sequences and promising performance is achieved.},
keywords={},
doi={10.1093/ietisy/e91-d.7.1848},
ISSN={1745-1361},
month={July},}
Copier
TY - JOUR
TI - View Invariant Human Action Recognition Based on Factorization and HMMs
T2 - IEICE TRANSACTIONS on Information
SP - 1848
EP - 1854
AU - Xi LI
AU - Kazuhiro FUKUI
PY - 2008
DO - 10.1093/ietisy/e91-d.7.1848
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E91-D
IS - 7
JA - IEICE TRANSACTIONS on Information
Y1 - July 2008
AB - This paper addresses the problem of view invariant action recognition using 2D trajectories of landmark points on human body. It is a challenging task since for a specific action category, the 2D observations of different instances might be extremely different due to varying viewpoint and changes in speed. By assuming that the execution of an action can be approximated by dynamic linear combination of a set of basis shapes, a novel view invariant human action recognition method is proposed based on non-rigid matrix factorization and Hidden Markov Models (HMMs). We show that the low dimensional weight coefficients of basis shapes by measurement matrix non-rigid factorization contain the key information for action recognition regardless of the viewpoint changing. Based on the extracted discriminative features, the HMMs is used for temporal dynamic modeling and robust action classification. The proposed method is tested using real life sequences and promising performance is achieved.
ER -