Training initialization of hidden markov models in human action recognition

Moghaddam, Z; Piccardi, M

Training initialization of hidden markov models in human action recognition

Moghaddam, Z Piccardi, M

Permalink

Publication Type:: Journal Article
Citation:: IEEE Transactions on Automation Science and Engineering, 2014, 11 (2), pp. 394 - 408
Issue Date:: 2014-01-01

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Text

Download Accepted Manuscript VersionText (11.75 kB)

Adobe PDF

Download Accepted ManuscriptAdobe PDF (762.58 kB)

View on publisher's site

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Moghaddam, Z	en_US
dc.contributor.author	Piccardi, M https://orcid.org/0000-0001-9250-6604	en_US
dc.date.issued	2014-01-01	en_US
dc.identifier.citation	IEEE Transactions on Automation Science and Engineering, 2014, 11 (2), pp. 394 - 408	en_US
dc.identifier.issn	1545-5955	en_US
dc.identifier.uri	http://hdl.handle.net/10453/35126
dc.description.abstract	Human action recognition in video is often approached by means of sequential probabilistic models as they offer a natural match to the temporal dimension of the actions. However, effective estimation of the models' parameters is critical if one wants to achieve significant recognition accuracy. Parameter estimation is typically performed over a set of training data by maximizing objective functions such as the data likelihood or the conditional likelihood. However, such functions are nonconvex in nature and subject to local maxima. This problem is major since any solution algorithm (expectation- maximization, gradient ascent, variational methods and others) requires an arbitrary initialization and can only find a corresponding local maximum. Exhaustive search is otherwise impossible since the number of local maxima is unknown. While no theoretical solutions are available for this problem, the only practicable mollification is to repeat training with different initializations until satisfactory cross-validation accuracy is attained. Such a process is overall empirical and highly time-consuming. In this paper, we propose two methods for one-off initialization of hidden Markov models achieving interesting tradeoffs between accuracy and training time. Experiments over three challenging human action video datasets (Weizmann, MuHAVi and Hollywood Human Actions) and with various feature sets measured from the frames (STIP descriptors, projection histograms, notable contour points) prove that the proposed one-off initializations are capable of achieving accuracy above the average of repeated random initializations and comparable to the best. In addition, the methods proposed are not restricted solely to human action recognition as they suit time series classification as a general problem. © 2004-2012 IEEE.	en_US
dc.relation.ispartof	IEEE Transactions on Automation Science and Engineering	en_US
dc.relation.isbasedon	10.1109/TASE.2013.2262940	en_US
dc.subject.classification	Industrial Engineering & Automation	en_US
dc.title	Training initialization of hidden markov models in human action recognition	en_US
dc.type	Journal Article
utslib.citation.volume	2	en_US
utslib.citation.volume	11	en_US
utslib.for	0906 Electrical and Electronic Engineering	en_US
utslib.for	0913 Mechanical Engineering	en_US
utslib.for	0910 Manufacturing Engineering	en_US
pubs.embargo.period	Not known	en_US
pubs.organisational-group	/University of Technology Sydney
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology/School of Electrical and Data Engineering
pubs.organisational-group	/University of Technology Sydney/Strength - GBDTC - Global Big Data Technologies
utslib.copyright.status	open_access
pubs.issue	2	en_US
pubs.publication-status	Published	en_US
pubs.volume	11	en_US

Abstract:

Human action recognition in video is often approached by means of sequential probabilistic models as they offer a natural match to the temporal dimension of the actions. However, effective estimation of the models' parameters is critical if one wants to achieve significant recognition accuracy. Parameter estimation is typically performed over a set of training data by maximizing objective functions such as the data likelihood or the conditional likelihood. However, such functions are nonconvex in nature and subject to local maxima. This problem is major since any solution algorithm (expectation- maximization, gradient ascent, variational methods and others) requires an arbitrary initialization and can only find a corresponding local maximum. Exhaustive search is otherwise impossible since the number of local maxima is unknown. While no theoretical solutions are available for this problem, the only practicable mollification is to repeat training with different initializations until satisfactory cross-validation accuracy is attained. Such a process is overall empirical and highly time-consuming. In this paper, we propose two methods for one-off initialization of hidden Markov models achieving interesting tradeoffs between accuracy and training time. Experiments over three challenging human action video datasets (Weizmann, MuHAVi and Hollywood Human Actions) and with various feature sets measured from the frames (STIP descriptors, projection histograms, notable contour points) prove that the proposed one-off initializations are capable of achieving accuracy above the average of repeated random initializations and comparable to the best. In addition, the methods proposed are not restricted solely to human action recognition as they suit time series classification as a general problem. © 2004-2012 IEEE.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/35126