Trajectory-based human action segmentation
Autor(a) principal: | |
---|---|
Data de Publicação: | 2015 |
Outros Autores: | , |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/10316/27816 https://doi.org/10.1016/j.patcog.2014.08.015 |
Resumo: | This paper proposes a sliding window approach, whose length and time shift are dynamically adaptable in order to improve model confidence, speed and segmentation accuracy in human action sequences. Activity recognition is the process of inferring an action class from a set of observations acquired by sensors. We address the temporal segmentation problem of body part trajectories in Cartesian Space in which features are generated using Discrete Fast Fourier Transform (DFFT) and Power Spectrum (PS). We pose this as an entropy minimization problem. Using entropy from the classifier output as a feedback parameter, we continuously adjust the two key parameters in a sliding window approach, to maximize the model confidence at every step. The proposed classifier is a Dynamic Bayesian Network (DBN) model where classes are estimated using Bayesian inference. We compare our approach with our previously developed fixed window method. Experiments show that our method accurately recognizes and segments activities, with improved model confidence and faster convergence times, exhibiting anticipatory capabilities. Our work demonstrates that entropy feedback mitigates variability problems, and our method is applicable in research areas where action segmentation and classification is used. A working demo source code is provided online for academical dissemination purposes, by requesting the authors. |
id |
RCAP_48d3341370c0f00132b8c566f58ed131 |
---|---|
oai_identifier_str |
oai:estudogeral.uc.pt:10316/27816 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
Trajectory-based human action segmentationMotion segmentationClassification frameworkSignal processingMotion variabilityAdaptive sliding windowThis paper proposes a sliding window approach, whose length and time shift are dynamically adaptable in order to improve model confidence, speed and segmentation accuracy in human action sequences. Activity recognition is the process of inferring an action class from a set of observations acquired by sensors. We address the temporal segmentation problem of body part trajectories in Cartesian Space in which features are generated using Discrete Fast Fourier Transform (DFFT) and Power Spectrum (PS). We pose this as an entropy minimization problem. Using entropy from the classifier output as a feedback parameter, we continuously adjust the two key parameters in a sliding window approach, to maximize the model confidence at every step. The proposed classifier is a Dynamic Bayesian Network (DBN) model where classes are estimated using Bayesian inference. We compare our approach with our previously developed fixed window method. Experiments show that our method accurately recognizes and segments activities, with improved model confidence and faster convergence times, exhibiting anticipatory capabilities. Our work demonstrates that entropy feedback mitigates variability problems, and our method is applicable in research areas where action segmentation and classification is used. A working demo source code is provided online for academical dissemination purposes, by requesting the authors.Elsevier2015-02info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/articlehttp://hdl.handle.net/10316/27816http://hdl.handle.net/10316/27816https://doi.org/10.1016/j.patcog.2014.08.015engSANTOS, Luís; KHOSHHAL, Kamrad; DIAS, Jorge - Trajectory-based human action segmentation. "Pattern Recognition". ISSN 0031-3203. Vol. 48 Nº. 2 (2015) p. 568–5790031-3203http://www.sciencedirect.com/science/article/pii/S003132031400329X#Santos, LuísKhoshhal, KamradDias, Jorgeinfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2020-05-25T12:20:24Zoai:estudogeral.uc.pt:10316/27816Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-19T20:53:46.562220Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
Trajectory-based human action segmentation |
title |
Trajectory-based human action segmentation |
spellingShingle |
Trajectory-based human action segmentation Santos, Luís Motion segmentation Classification framework Signal processing Motion variability Adaptive sliding window |
title_short |
Trajectory-based human action segmentation |
title_full |
Trajectory-based human action segmentation |
title_fullStr |
Trajectory-based human action segmentation |
title_full_unstemmed |
Trajectory-based human action segmentation |
title_sort |
Trajectory-based human action segmentation |
author |
Santos, Luís |
author_facet |
Santos, Luís Khoshhal, Kamrad Dias, Jorge |
author_role |
author |
author2 |
Khoshhal, Kamrad Dias, Jorge |
author2_role |
author author |
dc.contributor.author.fl_str_mv |
Santos, Luís Khoshhal, Kamrad Dias, Jorge |
dc.subject.por.fl_str_mv |
Motion segmentation Classification framework Signal processing Motion variability Adaptive sliding window |
topic |
Motion segmentation Classification framework Signal processing Motion variability Adaptive sliding window |
description |
This paper proposes a sliding window approach, whose length and time shift are dynamically adaptable in order to improve model confidence, speed and segmentation accuracy in human action sequences. Activity recognition is the process of inferring an action class from a set of observations acquired by sensors. We address the temporal segmentation problem of body part trajectories in Cartesian Space in which features are generated using Discrete Fast Fourier Transform (DFFT) and Power Spectrum (PS). We pose this as an entropy minimization problem. Using entropy from the classifier output as a feedback parameter, we continuously adjust the two key parameters in a sliding window approach, to maximize the model confidence at every step. The proposed classifier is a Dynamic Bayesian Network (DBN) model where classes are estimated using Bayesian inference. We compare our approach with our previously developed fixed window method. Experiments show that our method accurately recognizes and segments activities, with improved model confidence and faster convergence times, exhibiting anticipatory capabilities. Our work demonstrates that entropy feedback mitigates variability problems, and our method is applicable in research areas where action segmentation and classification is used. A working demo source code is provided online for academical dissemination purposes, by requesting the authors. |
publishDate |
2015 |
dc.date.none.fl_str_mv |
2015-02 |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/article |
format |
article |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/10316/27816 http://hdl.handle.net/10316/27816 https://doi.org/10.1016/j.patcog.2014.08.015 |
url |
http://hdl.handle.net/10316/27816 https://doi.org/10.1016/j.patcog.2014.08.015 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
SANTOS, Luís; KHOSHHAL, Kamrad; DIAS, Jorge - Trajectory-based human action segmentation. "Pattern Recognition". ISSN 0031-3203. Vol. 48 Nº. 2 (2015) p. 568–579 0031-3203 http://www.sciencedirect.com/science/article/pii/S003132031400329X# |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.publisher.none.fl_str_mv |
Elsevier |
publisher.none.fl_str_mv |
Elsevier |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799133823800180736 |