AdaBoost.MRF: Boosted Markov random forests and application to multilevel activity recognition

Tran The Truyen; Dinh Q. Phung; Hung H. Bui; Svetha Venkatesh

doi:10.1109/CVPR.2006.49

AdaBoost.MRF: Boosted Markov random forests and application to multilevel activity recognition

Tran The Truyen
, Dinh Q. Phung
, Hung H. Bui
, Svetha Venkatesh

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

16 Citations (Scopus)

Abstract

Activity recognition is an important issue in building intelligent monitoring systems. We address the recognition of multilevel activities in this paper via a conditional Markov random field (MRF), known as the dynamic conditional random field (DCRF). Parameter estimation in general MRFs using maximum likelihood is known to be computationally challenging (except for extreme cases), and thus we propose an efficient boosting-based algorithm AdaBoost.MRF for this task. Distinct from most existing work, our algorithm can handle hidden variables (missing labels) and is particularly attractive for smarthouse domains where reliable labels are often sparsely observed. Furthermore, our method works exclusively on trees and thus is guaranteed to converge. We apply the AdaBoost.MRF algorithm to a home video surveillance application and demonstrate its efficacy.

Original language	English
Title of host publication	Proceedings - 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2006
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Pages	1686-1693
Number of pages	8
ISBN (Print)	0769525970, 9780769525976
DOIs	https://doi.org/10.1109/CVPR.2006.49
Publication status	Published - 22 Dec 2006
Externally published	Yes
Event	IEEE Conference on Computer Vision and Pattern Recognition 2006 - New York, United States of America Duration: 17 Jun 2006 → 22 Jun 2006 https://ieeexplore.ieee.org/xpl/conhome/10924/proceeding?isnumber=34373 (Proceedings)

Conference

Conference	IEEE Conference on Computer Vision and Pattern Recognition 2006
Abbreviated title	CVPR 2006
Country/Territory	United States of America
City	New York
Period	17/06/06 → 22/06/06
Internet address	https://ieeexplore.ieee.org/xpl/conhome/10924/proceeding?isnumber=34373 (Proceedings)

Access to Document

10.1109/CVPR.2006.49

Cite this

Truyen, T. T., Phung, D. Q., Bui, H. H., & Venkatesh, S. (2006). AdaBoost.MRF: Boosted Markov random forests and application to multilevel activity recognition. In Proceedings - 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2006 (pp. 1686-1693). IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/CVPR.2006.49

@inproceedings{209b7231089f44a4860280776743484d,

title = "AdaBoost.MRF: Boosted Markov random forests and application to multilevel activity recognition",

abstract = "Activity recognition is an important issue in building intelligent monitoring systems. We address the recognition of multilevel activities in this paper via a conditional Markov random field (MRF), known as the dynamic conditional random field (DCRF). Parameter estimation in general MRFs using maximum likelihood is known to be computationally challenging (except for extreme cases), and thus we propose an efficient boosting-based algorithm AdaBoost.MRF for this task. Distinct from most existing work, our algorithm can handle hidden variables (missing labels) and is particularly attractive for smarthouse domains where reliable labels are often sparsely observed. Furthermore, our method works exclusively on trees and thus is guaranteed to converge. We apply the AdaBoost.MRF algorithm to a home video surveillance application and demonstrate its efficacy.",

author = "Truyen, \{Tran The\} and Phung, \{Dinh Q.\} and Bui, \{Hung H.\} and Svetha Venkatesh",

year = "2006",

month = dec,

day = "22",

doi = "10.1109/CVPR.2006.49",

language = "English",

isbn = "0769525970",

pages = "1686--1693",

booktitle = "Proceedings - 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2006",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

address = "United States of America",

note = "IEEE Conference on Computer Vision and Pattern Recognition 2006, CVPR 2006 ; Conference date: 17-06-2006 Through 22-06-2006",

url = "https://ieeexplore.ieee.org/xpl/conhome/10924/proceeding?isnumber=34373",

}

Truyen, TT, Phung, DQ, Bui, HH & Venkatesh, S 2006, AdaBoost.MRF: Boosted Markov random forests and application to multilevel activity recognition. in Proceedings - 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2006. IEEE, Institute of Electrical and Electronics Engineers, pp. 1686-1693, IEEE Conference on Computer Vision and Pattern Recognition 2006, New York, New York, United States of America, 17/06/06. https://doi.org/10.1109/CVPR.2006.49

AdaBoost.MRF: Boosted Markov random forests and application to multilevel activity recognition. / Truyen, Tran The; Phung, Dinh Q.; Bui, Hung H. et al.
Proceedings - 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2006. IEEE, Institute of Electrical and Electronics Engineers, 2006. p. 1686-1693.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - AdaBoost.MRF

T2 - IEEE Conference on Computer Vision and Pattern Recognition 2006

AU - Truyen, Tran The

AU - Phung, Dinh Q.

AU - Bui, Hung H.

AU - Venkatesh, Svetha

PY - 2006/12/22

Y1 - 2006/12/22

N2 - Activity recognition is an important issue in building intelligent monitoring systems. We address the recognition of multilevel activities in this paper via a conditional Markov random field (MRF), known as the dynamic conditional random field (DCRF). Parameter estimation in general MRFs using maximum likelihood is known to be computationally challenging (except for extreme cases), and thus we propose an efficient boosting-based algorithm AdaBoost.MRF for this task. Distinct from most existing work, our algorithm can handle hidden variables (missing labels) and is particularly attractive for smarthouse domains where reliable labels are often sparsely observed. Furthermore, our method works exclusively on trees and thus is guaranteed to converge. We apply the AdaBoost.MRF algorithm to a home video surveillance application and demonstrate its efficacy.

AB - Activity recognition is an important issue in building intelligent monitoring systems. We address the recognition of multilevel activities in this paper via a conditional Markov random field (MRF), known as the dynamic conditional random field (DCRF). Parameter estimation in general MRFs using maximum likelihood is known to be computationally challenging (except for extreme cases), and thus we propose an efficient boosting-based algorithm AdaBoost.MRF for this task. Distinct from most existing work, our algorithm can handle hidden variables (missing labels) and is particularly attractive for smarthouse domains where reliable labels are often sparsely observed. Furthermore, our method works exclusively on trees and thus is guaranteed to converge. We apply the AdaBoost.MRF algorithm to a home video surveillance application and demonstrate its efficacy.

UR - https://www.scopus.com/pages/publications/33845575878

U2 - 10.1109/CVPR.2006.49

DO - 10.1109/CVPR.2006.49

M3 - Conference Paper

AN - SCOPUS:33845575878

SN - 0769525970

SN - 9780769525976

SP - 1686

EP - 1693

BT - Proceedings - 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2006

PB - IEEE, Institute of Electrical and Electronics Engineers

Y2 - 17 June 2006 through 22 June 2006

ER -

AdaBoost.MRF: Boosted Markov random forests and application to multilevel activity recognition

Abstract

Conference

Access to Document

Other files and links

Cite this