Neural tensor network and adaptive graph convolution for sports action recognition

Lin  Teng; Hang  Li; Yuchang  Si

doi:10.6180/jase.202606_29(6).0015

Neural tensor network and adaptive graph convolution for sports action recognition

Lin Teng, Hang LiThis email address is being protected from spambots. You need JavaScript enabled to view it., and Yuchang SiThis email address is being protected from spambots. You need JavaScript enabled to view it.

College of Artificial Intelligence, Shenyang Normal University, Shenyang 110034, 7 China

Received: June 14, 2025
Accepted: October 6, 2025
Publication Date: November 12, 2025

Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Download Citation: ||https://doi.org/10.6180/jase.202606_29(6).0015

Current human bone action recognition algorithms have some problems such as insufficiently detailed de scription of the global relationship and insufficient mining of spatio-temporal features. Therefore, this paper proposes a novel sports action recognition based on neural tensor network and adaptive graph convolution. Firstly, the attention mechanism and neural tensor network (NTN) algorithm are used to solve the connection strength between each pair of joint nodes and construct the global adjacency matrix. Secondly, by using the topK strategy, the topK neighbor nodes are dynamically selected based on the connection strength to update the global adjacency matrix. Thirdly, the hybrid pooling model is adopted to extract the global context information and the temporal key frame features. By simultaneously modeling joint information, bone information, joint movement information and bone movement information, the representation ability of the features extracted by the model for movements is strengthened. The experimental results on the Something-Something V1&V2 and Kinetics-400 datasets show that the proposed model in this paper outperforms most other advanced action recognition methods, proving that this new model can effectively improve the performance of action recognition.

Keywords: sports action recognition, neural tensor network, adaptive graph convolution, topK strategy, hybrid pooling model

[1] N.Manakitsa,G.S.Maraslidis,L.Moysis,andG.F. Fragulis, (2024) “Areviewofmachine learning and deep learning for object detection, semantic segmentation, and humanaction recognition in machine and roboticvision"Technologies12(2):15.DOI:10.3390/technologies12020015.
[2] D.Guo,K.Li,B.Hu,Y.Zhang,andM.Wang,(2024) “Benchmarkingmicro-actionrecognition:Dataset,methods, and applications" IEEE Transactions on Circuits and SystemsforVideoTechnology34(7):6238–6252. DOI:10.1109/TCSVT.2024.3358415.
[3] S. Yin, H. Li, A. A. Laghari, T. R. Gadekallu, G. A. Sampedro, andA.Almadhor, (2024)“Ananomaly detection model based on deep auto-encoder and capsule graph convolution via sparrow search algorithm in 6G Internet of Everything "IEEE Internet of Things Journal 11(18):29402–29411.DOI:10.1109/JIOT.2024.3353337.
[4] M.AntounandD.Asmar,(2023)“Humanobjectinter action detection: Design and survey "Image and Vision Computing130:104617.DOI:10.1016/j.imavis.2022.104617.
[5] H.Zhou,W.Zhou,Y.Zhou,andH.Li,(2021)“Spatial temporal multi-cuenet work for signlanguage recognition and translation"IEEE Transactionson Multimedia 24:768–779.DOI:10.1109/TMM.2021.3059098.
[6] P.Ravbar,K.Branson,andJ.H.Simpson,(2019)“An automatic behavior recognition system classifies animal behaviors using movements and their temporal context" Journalofneurosciencemethods326:108352.DOI: 10.1016/j.jneumeth.2019.108352.
[7] L. Pigou, A. Van Den Oord, S. Dieleman, M. Van Herreweghe,andJ.Dambre,(2018)“Beyondtemporal pooling: Recurrence and temporal convolutions for gesture recognition in video" International Journal of Computer Vision 126(2): 430–439. DOI: 10.1007/s11263 016-0957-7.
[8] Y. Jiang and S. Yin, (2023) “Heterogenous-view occluded expression data recognition based on cycle-consistent adversarial network and K-SVD dictionary learning under intelligent cooperative robot environment" Computer Science and Information Systems 20(4): 1869–1883. DOI: 10.2298/CSIS221228034J.
[9] Y. Zhou, X. Yan, Z.-Q. Cheng, Y. Yan, Q. Dai, and X.-S. Hua. “Blockgcn: Redefine topology awareness for skeleton-based action recognition”. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2024, 2049–2058. DOI: 10.1109/CVPR52733.2024.00200.
[10] Y. Abbas and A. Jalal. “Drone-based human action recognition for surveillance: a multi-feature ap proach”. In: 2024 International Conference on Engineering & Computing Technologies (ICECT). IEEE. 2024, 1 6. DOI: 10.1109/ICECT61618.2024.10581378.
[11] Y. Ma and R. Wang, (2024) “Relative-position embed ding based spatially and temporally decoupled Trans former for action recognition" Pattern Recognition 145: 109905. DOI: 10.1016/j.patcog.2023.109905.
[12] M. Munsif, N. Khan, A. Hussain, M. J. Kim, and S. W. Baik, (2024) “Darkness-adaptive action recognition: Leveraging efficient tubelet slow-fast network for industrial applications" IEEE Transactions on Indus trial Informatics 20(12): 13676–13686. DOI: 10.1109/ TII.2024.3431070.
[13] I. A. Abro and A. Jalal. “Multi-Modal Sensors Fusion for Fall Detection and Action Recognition in Indoor Environment”. In: 2024 3rd International Conference on Emerging Trends in Electrical, Control, and Telecommu nication Engineering (ETECTE). IEEE. 2024, 1–6. DOI: 10.1109/ETECTE63967.2024.10823705.
[14] N. Siddiqui, P. Tirupattur, and M. Shah. “DVANet: Disentangling view and action features for multi view action recognition”. In: Proceedings of the AAAI Conference on Artificial Intelligence. 38. 5. 2024, 4873 4881. DOI: 10.1609/aaai.v38i5.28290.
[15] S. Rakshit, T. Davies, M. M. Moradi, G. McSwiggan, G. Nair, J. Mateu, andA.Baddeley, (2019) “Fast kernel smoothing of point patterns on a large network using two-dimensional convolution" International Statistical Review 87(3): 531–556. DOI: 10.1111/insr.12327.
[16] X. Ma, W. Xu, H.Guan, and X. Zhang, (2024) “Three dimensional image recognition of soybean canopy based on improved multi-view network" Industrial Crops and Products 222: 119544. DOI: 10.1016/j.indcrop.2024. 119544.
[17] Y. Zhang, J. Li, N. Jiang, G. Wu, H. Zhang, Z. Shi, Z. Liu, Z. Wu, and X. Liu, (2023) “Temporal transformer networks with self-supervision for action recognition" IEEE Internet of Things Journal 10(14): 12999–13011. DOI: 10.1109/JIOT.2023.3257992.
[18] S. Basak, P. Corcoran, R. McDonnell, and M. Schukat, (2022) “3D face-model reconstruction from a single image: A feature aggregation approach using hierarchical trans former with weak supervision" Neural Networks 156: 108–122. DOI: 10.1016/j.neunet.2022.09.019.
[19] H.QiuandB.Hou,(2024) “Multi-grained clip focus for skeleton-based action recognition" Pattern Recognition 148: 110188. DOI: 10.1016/j.patcog.2023.110188.
[20] M.Batool, M. Alotaibi, S. R. Alotaibi, D. A. Al Hammadi, M. A. Jamal, A. Jalal, and B. Lee, (2024) “Multimodal human action recognition framework using an improved CNNGRU classifier" IEEE Access 12: 158388 158406. DOI: 10.1109/ACCESS.2024.3481631.
[21] K. Peng, C. Yin, J. Zheng, R. Liu, D. Schneider, J. Zhang, K. Yang, M. S. Sarfraz, R. Stiefelhagen, and A. Roitberg. “Navigating open set scenarios for skeleton-based action recognition”. In: Proceedings of the AAAI conference on artificial intelligence. 38. 5. 2024, 4487–4496. DOI: 10.1609/aaai.v38i5.28247.
[22] R. Xie, Y. Jiang, and J. Yu. “Two-stream adaptive graph convolutional network with multi-head attention mechanism for industrial safety detection”. In: Fifth International Conference on Telecommunications, Optics, and Computer Science (TOCS 2024). 13629. SPIE. 2025, 710–716. DOI: 10.1117/12.3067995.
[23] Z. Bai, Q. Ding, H. Xu, J. Chi, X. Zhang, and T. Sun, (2022) “Skeleton-based similar action recognition through integrating the salient image feature into a center connected graph convolutional network" Neurocomputing 507: 40–53. DOI: 10.1016/j.neucom.2022.07.080.
[24] W. Hong, W. Xu, J. Qi, and Y. Weng, (2019) “Neural tensor network for multi-label classification" IEEE Ac cess 7: 96936–96941. DOI: 10.1109/ACCESS.2019.2930206.
[25] L.Teng,Y. Qiao, M. Shafiq, G. Srivastava, A. R. Javed, T. R. Gadekallu, and S. Yin, (2023) “FLPK-BiSeNet: Federated learning based on priori knowledge and bilateral segmentation network for image edge extraction" IEEE Transactions on Network and Service Management 20(2): 1529–1542. DOI: 10.1109/TNSM.2023.3273991.
[26] S. Jang, H. Lee, W. J. Kim, J. Lee, S. Woo, and S. Lee, (2024) “Multi-scale structural graph convolutional net work for skeleton-based action recognition" IEEE Trans actions on Circuits and Systems for Video Technology 34(8): 7244–7258. DOI: 10.1109/TCSVT.2024.3375512.
[27] E. Dastbaravardeh, S. Askarpour, M. Saberi Anari, and K. Rezaee, (2024) “Channel attention-based approach with autoencoder network for human action recognition in low-resolution frames" International Journal of Intelligent Systems 2024(1): 1052344. DOI: 10.1155/2024/1052344.
[28] H. Xu, Y. Gao, Z. Hui, J. Li, and X. Gao, (2025) “Language knowledge-assisted representation learning for skeleton-based action recognition" IEEE Transactions on Multimedia 27: 5784–5799. DOI: 10.1109/TMM. 2025.3543034.
[29] S.B.Khobdeh,M.R.Yamaghani,andS.K.Sareshkeh, (2024) “Basketball action recognition based on the com bination of YOLO and a deep fuzzy LSTM network: SB Khobdeh et al." The Journal of Supercomputing 80(3): 3528–3553. DOI: 10.1007/s11227-023-05611-7.
[30] A. O. Kolawole, M. E. Irhebhude, and P. O. Odion, (2025) “Human Action Recognition in Military Obstacle Crossing Using HOG and Region-Based Descriptors" Journal of Computing Theories and Applications 2(3): 410–426. DOI: 10.62411/jcta.12195.
[31] M. A. Khan, K. Javed, S. A. Khan, T. Saba, U. Habib, J. A. Khan, and A. A. Abbasi, (2024) “Human action recognition using fusion of multiview and deep features: an application to video surveillance" Multimedia tools and applications 83(5): 14885–14911. DOI: 10.1007/s11042-020-08806-9.
[32] A. C. Cob-Parro, C. Losada-Gutiérrez, M. Marrón Romera, A. Gardel-Vicente, and I. Bravo-Munoz, (2024) “A new framework for deep learning video based Human Action Recognition on the edge" Expert Systems with Applications 238: 122220. DOI: 10.1016/j.eswa. 2023.122220.
[33] X. Wang, S. Zhang, J. Cen, C. Gao, Y. Zhang, D. Zhao, and N. Sang, (2024) “Clip-guided prototype modulating for few-shot action recognition" International Journal of Computer Vision 132(6): 1899–1912. DOI: 10.1007/s11263-023-01917-4.

Latest Articles