Human Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework

Human action recognition (HAR) by skeleton data is considered a potential research aspect in computer vision. Three-dimensional HAR with skeleton data has been used commonly because of its effective and efficient results. Several models have been developed for learning spatiotemporal parameters from...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Faisal Mehmood, Enqing Chen, Muhammad Azeem Akbar, Abeer Abdulaziz Alsanad
Formato: article
Lenguaje:EN
Publicado: MDPI AG 2021
Materias:
Acceso en línea:https://doaj.org/article/ef9865d76f654905b711d50b6b7c2dd9
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:ef9865d76f654905b711d50b6b7c2dd9
record_format dspace
spelling oai:doaj.org-article:ef9865d76f654905b711d50b6b7c2dd92021-11-11T15:42:09ZHuman Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework10.3390/electronics102127082079-9292https://doaj.org/article/ef9865d76f654905b711d50b6b7c2dd92021-11-01T00:00:00Zhttps://www.mdpi.com/2079-9292/10/21/2708https://doaj.org/toc/2079-9292Human action recognition (HAR) by skeleton data is considered a potential research aspect in computer vision. Three-dimensional HAR with skeleton data has been used commonly because of its effective and efficient results. Several models have been developed for learning spatiotemporal parameters from skeleton sequences. However, two critical problems exist: (1) previous skeleton sequences were created by connecting different joints with a static order; (2) earlier methods were not efficient enough to focus on valuable joints. Specifically, this study aimed to (1) demonstrate the ability of convolutional neural networks to learn spatiotemporal parameters of skeleton sequences from different frames of human action, and (2) to combine the process of all frames created by different human actions and fit in the spatial structure information necessary for action recognition, using multi-task learning networks (MTLNs). The results were significantly improved compared with existing models by executing the proposed model on an NTU RGB+D dataset, an SYSU dataset, and an SBU Kinetic Interaction dataset. We further implemented our model on noisy expected poses from subgroups of the Kinetics dataset and the UCF101 dataset. The experimental results also showed significant improvement using our proposed model.Faisal MehmoodEnqing ChenMuhammad Azeem AkbarAbeer Abdulaziz AlsanadMDPI AGarticlehuman action recognition (HAR)skeleton dataspatiotemporalmulti-task learning network (MTLN)convolutional neural network (CNN)ElectronicsTK7800-8360ENElectronics, Vol 10, Iss 2708, p 2708 (2021)
institution DOAJ
collection DOAJ
language EN
topic human action recognition (HAR)
skeleton data
spatiotemporal
multi-task learning network (MTLN)
convolutional neural network (CNN)
Electronics
TK7800-8360
spellingShingle human action recognition (HAR)
skeleton data
spatiotemporal
multi-task learning network (MTLN)
convolutional neural network (CNN)
Electronics
TK7800-8360
Faisal Mehmood
Enqing Chen
Muhammad Azeem Akbar
Abeer Abdulaziz Alsanad
Human Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework
description Human action recognition (HAR) by skeleton data is considered a potential research aspect in computer vision. Three-dimensional HAR with skeleton data has been used commonly because of its effective and efficient results. Several models have been developed for learning spatiotemporal parameters from skeleton sequences. However, two critical problems exist: (1) previous skeleton sequences were created by connecting different joints with a static order; (2) earlier methods were not efficient enough to focus on valuable joints. Specifically, this study aimed to (1) demonstrate the ability of convolutional neural networks to learn spatiotemporal parameters of skeleton sequences from different frames of human action, and (2) to combine the process of all frames created by different human actions and fit in the spatial structure information necessary for action recognition, using multi-task learning networks (MTLNs). The results were significantly improved compared with existing models by executing the proposed model on an NTU RGB+D dataset, an SYSU dataset, and an SBU Kinetic Interaction dataset. We further implemented our model on noisy expected poses from subgroups of the Kinetics dataset and the UCF101 dataset. The experimental results also showed significant improvement using our proposed model.
format article
author Faisal Mehmood
Enqing Chen
Muhammad Azeem Akbar
Abeer Abdulaziz Alsanad
author_facet Faisal Mehmood
Enqing Chen
Muhammad Azeem Akbar
Abeer Abdulaziz Alsanad
author_sort Faisal Mehmood
title Human Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework
title_short Human Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework
title_full Human Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework
title_fullStr Human Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework
title_full_unstemmed Human Action Recognition of Spatiotemporal Parameters for Skeleton Sequences Using MTLN Feature Learning Framework
title_sort human action recognition of spatiotemporal parameters for skeleton sequences using mtln feature learning framework
publisher MDPI AG
publishDate 2021
url https://doaj.org/article/ef9865d76f654905b711d50b6b7c2dd9
work_keys_str_mv AT faisalmehmood humanactionrecognitionofspatiotemporalparametersforskeletonsequencesusingmtlnfeaturelearningframework
AT enqingchen humanactionrecognitionofspatiotemporalparametersforskeletonsequencesusingmtlnfeaturelearningframework
AT muhammadazeemakbar humanactionrecognitionofspatiotemporalparametersforskeletonsequencesusingmtlnfeaturelearningframework
AT abeerabdulazizalsanad humanactionrecognitionofspatiotemporalparametersforskeletonsequencesusingmtlnfeaturelearningframework
_version_ 1718434122662150144