Publication
MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition
| Summary: | Posture recognition is critical in modern educational and office environments for preventing musculoskeletal disorders and maintaining cognitive performance. Existing methods based on human keypoint detection typically rely on convolutional neural networks (CNNs) and single-scale features, which limit representation capacity and suffer from overfitting under small-sample conditions. To address these issues, we propose MSBN-SPose, a Multi-Scale Bayesian Neuro-Symbolic Posture Recognition framework that integrates geometric features at multiple levels—including global body structure, local regions, facial landmarks, distances, and angles—extracted from OpenPose keypoints. These features are processed by a multi-branch Bayesian neural architecture that models epistemic uncertainty, enabling improved generalization and robustness. Furthermore, a lightweight neuro-symbolic reasoning module incorporates human-understandable rules into the inference process, enhancing transparency and interpretability. To support real-world evaluation, we construct the USSP dataset, a diverse, classroom-representative collection of student postures under varying conditions. Experimental results show that MSBN-SPose achieves 96.01% accuracy on USSP, outperforming baseline and traditional methods under data-limited scenarios. |
|---|---|
| Main Authors: | Tavares, Adriano |
| Other Authors: | Carlos Lima; Lima, Carlos; Gomes, Tiago Manuel Ribeiro; Liang, Yanchun |
| Subject: | posture recognition convolutional neural network (CNN) multi-scale bayesian neural network (BNN) hierarchical neuro-symbolic method Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| Year: | 2025 |
| Country: | Portugal |
| Document type: | article |
| Access type: | open access |
| Associated institution: | Universidade do Minho |
| Language: | English |
| Origin: | RepositóriUM - Universidade do Minho |
| _version_ | 1877775999987351552 |
|---|---|
| author | Tavares, Adriano |
| author2 | Carlos Lima Lima, Carlos Gomes, Tiago Manuel Ribeiro Liang, Yanchun |
| author2_role | author author author author |
| author_facet | Tavares, Adriano Carlos Lima Lima, Carlos Gomes, Tiago Manuel Ribeiro Liang, Yanchun |
| author_role | author |
| contributor_name_str_mv | Universidade do Minho |
| country_str | PT |
| creators_json_txt | [{\"Person.name\":\"Tavares, Adriano\"},{\"Person.name\":\"Carlos Lima\"},{\"Person.name\":\"Lima, Carlos\"},{\"Person.name\":\"Gomes, Tiago Manuel Ribeiro\"},{\"Person.name\":\"Liang, Yanchun\"}] |
| datacite.contributors.contributor.contributorName.fl_str_mv | Universidade do Minho |
| datacite.creators.creator.creatorName.fl_str_mv | Tavares, Adriano Carlos Lima Lima, Carlos Gomes, Tiago Manuel Ribeiro Liang, Yanchun |
| datacite.date.Accepted.fl_str_mv | 2025-09-30T00:00:00Z |
| datacite.rights.fl_str_mv | http://purl.org/coar/access_right/c_abf2 |
| datacite.subjects.subject.fl_str_mv | posture recognition convolutional neural network (CNN) multi-scale bayesian neural network (BNN) hierarchical neuro-symbolic method Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| datacite.titles.title.fl_str_mv | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| dc.contributor.none.fl_str_mv | Universidade do Minho |
| dc.creator.none.fl_str_mv | Tavares, Adriano Carlos Lima Lima, Carlos Gomes, Tiago Manuel Ribeiro Liang, Yanchun |
| dc.date.Accepted.fl_str_mv | 2025-09-30T00:00:00Z |
| dc.format.none.fl_str_mv | application/pdf |
| dc.identifier.none.fl_str_mv | https://hdl.handle.net/1822/97359 |
| dc.language.none.fl_str_mv | eng |
| dc.publisher.none.fl_str_mv | MDPI |
| dc.rights.cclincense.fl_str_mv | http://creativecommons.org/licenses/by/4.0/ |
| dc.rights.none.fl_str_mv | http://purl.org/coar/access_right/c_abf2 |
| dc.rights.rights.copyright.fl_str_mv | openAccess |
| dc.subject.none.fl_str_mv | posture recognition convolutional neural network (CNN) multi-scale bayesian neural network (BNN) hierarchical neuro-symbolic method Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| dc.title.fl_str_mv | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| dc.type.none.fl_str_mv | http://purl.org/coar/resource_type/c_6501 |
| description | Posture recognition is critical in modern educational and office environments for preventing musculoskeletal disorders and maintaining cognitive performance. Existing methods based on human keypoint detection typically rely on convolutional neural networks (CNNs) and single-scale features, which limit representation capacity and suffer from overfitting under small-sample conditions. To address these issues, we propose MSBN-SPose, a Multi-Scale Bayesian Neuro-Symbolic Posture Recognition framework that integrates geometric features at multiple levels—including global body structure, local regions, facial landmarks, distances, and angles—extracted from OpenPose keypoints. These features are processed by a multi-branch Bayesian neural architecture that models epistemic uncertainty, enabling improved generalization and robustness. Furthermore, a lightweight neuro-symbolic reasoning module incorporates human-understandable rules into the inference process, enhancing transparency and interpretability. To support real-world evaluation, we construct the USSP dataset, a diverse, classroom-representative collection of student postures under varying conditions. Experimental results show that MSBN-SPose achieves 96.01% accuracy on USSP, outperforming baseline and traditional methods under data-limited scenarios. |
| dirty | 0 |
| eu_rights_str_mv | openAccess |
| format | article |
| id | rum_68aed2bb757efa2d6ce455e7c0975cc1 |
| identifier.url.fl_str_mv | https://hdl.handle.net/1822/97359 |
| inst_facet_str | urn:organizationAcronym:repositorium{{{_:::_}}}Universidade do Minho |
| instacron_str | repositorium |
| institution | Universidade do Minho |
| instname_str | Universidade do Minho |
| language | eng |
| network_acronym_str | rum |
| network_name_str | RepositóriUM - Universidade do Minho |
| oai_identifier_str | oai:repositorium.uminho.pt:1822/97359 |
| organization_str_mv | urn:organizationAcronym:repositorium |
| person_str_mv | Tavares, Adriano Carlos Lima Lima, Carlos Gomes, Tiago Manuel Ribeiro Liang, Yanchun |
| publishDate | 2025 |
| publisher.none.fl_str_mv | MDPI |
| repo_facet_str | urn:repositoryAcronym:rum{{{_:::_}}}RepositóriUM - Universidade do Minho |
| reponame_str | RepositóriUM - Universidade do Minho |
| repository_id_str | urn:repositoryAcronym:rum |
| service_str_mv | urn:repositoryAcronym:rum |
| spelling | engMDPIengPosture recognition is critical in modern educational and office environments for preventing musculoskeletal disorders and maintaining cognitive performance. Existing methods based on human keypoint detection typically rely on convolutional neural networks (CNNs) and single-scale features, which limit representation capacity and suffer from overfitting under small-sample conditions. To address these issues, we propose MSBN-SPose, a Multi-Scale Bayesian Neuro-Symbolic Posture Recognition framework that integrates geometric features at multiple levels—including global body structure, local regions, facial landmarks, distances, and angles—extracted from OpenPose keypoints. These features are processed by a multi-branch Bayesian neural architecture that models epistemic uncertainty, enabling improved generalization and robustness. Furthermore, a lightweight neuro-symbolic reasoning module incorporates human-understandable rules into the inference process, enhancing transparency and interpretability. To support real-world evaluation, we construct the USSP dataset, a diverse, classroom-representative collection of student postures under varying conditions. Experimental results show that MSBN-SPose achieves 96.01% accuracy on USSP, outperforming baseline and traditional methods under data-limited scenarios.application/pdfengMSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognitionTavares, AdrianoCarlos LimaLima, CarlosGomes, Tiago Manuel RibeiroLiang, YanchunHostingInstitutionOrganizationalUniversidade do Minhoe-mailmailto:repositorium@usdb.uminho.ptrepositorium@usdb.uminho.ptCITATIONWang, S., Tavares, A., Lima, C., Gomes, T., Zhang, Y., & Liang, Y. (2025). MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic Approach for Sitting Posture Recognition. Electronics, 14(19), 3889. https://doi.org/10.3390/electronics14193889EISSNIsPartOf2079-9292DOIIsPartOf10.3390/electronics14193889URLHasVersionhttps://www.mdpi.com/2079-9292/14/19/38892025-09-302025-09-30T00:00:00ZHandlehttps://hdl.handle.net/1822/97359http://purl.org/coar/access_right/c_abf2open accessposture recognitionconvolutional neural network (CNN)multi-scalebayesian neural network (BNN)hierarchical neuro-symbolic methodhttp://www.oecd.org/science/inno/38235147.pdfFields of Science and Technology (FOS)Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática3624268 bytesliteraturehttp://purl.org/coar/resource_type/c_6501journal article2025-09-30http://creativecommons.org/licenses/by/4.0/openAccesshttp://purl.org/coar/access_right/c_abf2application/pdfhttps://repositorium.uminho.pt/bitstreams/92250f21-c1ce-4d4f-a8db-87b1c6bf7ef2/download1418 |
| spellingShingle | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition Tavares, Adriano posture recognition convolutional neural network (CNN) multi-scale bayesian neural network (BNN) hierarchical neuro-symbolic method Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| status | SINGLETON |
| subject.fl_str_mv | posture recognition convolutional neural network (CNN) multi-scale bayesian neural network (BNN) hierarchical neuro-symbolic method |
| subject.other.fl_str_mv | Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| title | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| title_full | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| title_fullStr | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| title_full_unstemmed | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| title_short | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| title_sort | MSBN-SPose: A Multi-Scale Bayesian Neuro-Symbolic approach for sitting posture recognition |
| topic | posture recognition convolutional neural network (CNN) multi-scale bayesian neural network (BNN) hierarchical neuro-symbolic method Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| topic_facet | posture recognition convolutional neural network (CNN) multi-scale bayesian neural network (BNN) hierarchical neuro-symbolic method Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| url | https://hdl.handle.net/1822/97359 |
| visible | 1 |
Funded activities
No funded projects found.