Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognition
IntroductionAccurate recognition of martial arts leg poses is essential for applications in sports analytics, rehabilitation, and human-computer interaction. Traditional pose recognition models, relying on sequential or convolutional approaches, often struggle to capture the complex spatial-temporal...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-01-01
|
| Series: | Frontiers in Neurorobotics |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnbot.2024.1520983/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832592129229062144 |
|---|---|
| author | Shun Yao Yihan Ping Xiaoyu Yue Xiaoyu Yue He Chen He Chen |
| author_facet | Shun Yao Yihan Ping Xiaoyu Yue Xiaoyu Yue He Chen He Chen |
| author_sort | Shun Yao |
| collection | DOAJ |
| description | IntroductionAccurate recognition of martial arts leg poses is essential for applications in sports analytics, rehabilitation, and human-computer interaction. Traditional pose recognition models, relying on sequential or convolutional approaches, often struggle to capture the complex spatial-temporal dependencies inherent in martial arts movements. These methods lack the ability to effectively model the nuanced dynamics of joint interactions and temporal progression, leading to limited generalization in recognizing complex actions.MethodsTo address these challenges, we propose PoseGCN, a Graph Convolutional Network (GCN)-based model that integrates spatial, temporal, and contextual features through a novel framework. PoseGCN leverages spatial-temporal graph encoding to capture joint motion dynamics, an action-specific attention mechanism to assign importance to relevant joints depending on the action context, and a self-supervised pretext task to enhance temporal robustness and continuity. Experimental results on four benchmark datasets—Kinetics-700, Human3.6M, NTU RGB+D, and UTD-MHAD—demonstrate that PoseGCN outperforms existing models, achieving state-of-the-art accuracy and F1 scores.Results and discussionThese findings highlight the model's capacity to generalize across diverse datasets and capture fine-grained pose details, showcasing its potential in advancing complex pose recognition tasks. The proposed framework offers a robust solution for precise action recognition and paves the way for future developments in multi-modal pose analysis. |
| format | Article |
| id | doaj-art-29637558e8db4b10a80adb82959ef835 |
| institution | Kabale University |
| issn | 1662-5218 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neurorobotics |
| spelling | doaj-art-29637558e8db4b10a80adb82959ef8352025-01-21T14:19:18ZengFrontiers Media S.A.Frontiers in Neurorobotics1662-52182025-01-011810.3389/fnbot.2024.15209831520983Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognitionShun Yao0Yihan Ping1Xiaoyu Yue2Xiaoyu Yue3He Chen4He Chen5Department of Public Instruction, ChangJiang Polytechnic of Art and Engineering, Jingzhou, ChinaSchool of Computer Science, Northwestern University, Evanston, IL, United StatesSchool of Physical Education, Hubei University of Science and Technology, Xianning, ChinaCollege of Physical Education, Sangmyung University, Seoul, Republic of KoreaSchool of Physical Education, Hubei University of Science and Technology, Xianning, ChinaCollege of Physical Education, Sangmyung University, Seoul, Republic of KoreaIntroductionAccurate recognition of martial arts leg poses is essential for applications in sports analytics, rehabilitation, and human-computer interaction. Traditional pose recognition models, relying on sequential or convolutional approaches, often struggle to capture the complex spatial-temporal dependencies inherent in martial arts movements. These methods lack the ability to effectively model the nuanced dynamics of joint interactions and temporal progression, leading to limited generalization in recognizing complex actions.MethodsTo address these challenges, we propose PoseGCN, a Graph Convolutional Network (GCN)-based model that integrates spatial, temporal, and contextual features through a novel framework. PoseGCN leverages spatial-temporal graph encoding to capture joint motion dynamics, an action-specific attention mechanism to assign importance to relevant joints depending on the action context, and a self-supervised pretext task to enhance temporal robustness and continuity. Experimental results on four benchmark datasets—Kinetics-700, Human3.6M, NTU RGB+D, and UTD-MHAD—demonstrate that PoseGCN outperforms existing models, achieving state-of-the-art accuracy and F1 scores.Results and discussionThese findings highlight the model's capacity to generalize across diverse datasets and capture fine-grained pose details, showcasing its potential in advancing complex pose recognition tasks. The proposed framework offers a robust solution for precise action recognition and paves the way for future developments in multi-modal pose analysis.https://www.frontiersin.org/articles/10.3389/fnbot.2024.1520983/fullmartial arts pose recognitionspatial-temporal graph encodingGraph Convolutional Networksaction-specific attentionself-supervised learning |
| spellingShingle | Shun Yao Yihan Ping Xiaoyu Yue Xiaoyu Yue He Chen He Chen Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognition Frontiers in Neurorobotics martial arts pose recognition spatial-temporal graph encoding Graph Convolutional Networks action-specific attention self-supervised learning |
| title | Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognition |
| title_full | Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognition |
| title_fullStr | Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognition |
| title_full_unstemmed | Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognition |
| title_short | Graph Convolutional Networks for multi-modal robotic martial arts leg pose recognition |
| title_sort | graph convolutional networks for multi modal robotic martial arts leg pose recognition |
| topic | martial arts pose recognition spatial-temporal graph encoding Graph Convolutional Networks action-specific attention self-supervised learning |
| url | https://www.frontiersin.org/articles/10.3389/fnbot.2024.1520983/full |
| work_keys_str_mv | AT shunyao graphconvolutionalnetworksformultimodalroboticmartialartslegposerecognition AT yihanping graphconvolutionalnetworksformultimodalroboticmartialartslegposerecognition AT xiaoyuyue graphconvolutionalnetworksformultimodalroboticmartialartslegposerecognition AT xiaoyuyue graphconvolutionalnetworksformultimodalroboticmartialartslegposerecognition AT hechen graphconvolutionalnetworksformultimodalroboticmartialartslegposerecognition AT hechen graphconvolutionalnetworksformultimodalroboticmartialartslegposerecognition |