FOREX Press I. J. of Electrical & Electronics Research
Support Open Access
  • Home/
  • IJEER-130414

Review Article |

Cascaded Segmentation-Classification Framework Optimized for Sports Action Recognition Using Still Images: A Comparative Analysis of PSO and Grid Search

Author(s): Audrey Huong (PhD)1*, Ser Lee Loh (PhD)2, Kok Beng Gan (PhD)3, and Xavier Ngu (PhD)4

Published In : International Journal of Electrical and Electronics Research (IJEER)        Volume 13, Issue 4

Publisher : FOREX Publication

Published : 10 December 2025

e-ISSN : 2347-470X

Page(s) : 735-748

DOI: https://doi.org/10.37391/IJEER.130414

Abstract

The conventional method of evaluating human actions and activities requires integrating complex hardware and interpretation systems. This article proposes using a hybrid segmentation-classification system to recognize human sports action automatically based on still frames. This study tested the proposed framework on a public dataset containing images of humans performing three athletic actions: dancing, performing martial arts, and playing net sports. This framework combined a two-stage U-Net and EfficientNet-B0, and GoogleNet, and is optimized using particle swarm optimization (PSO) and grid search for cascaded segmentation and classification problems. Results indicated that PSO improves segmentation and classification accuracies by 5 % and 60 %, respectively, compared to the conventional grid search. The PSO segmentation results showed good agreement between the predicted mask and its ground truth, with overlapping scores of 0.85-0.9. The consecutive classification experiments revealed a slight superiority in the performance of EfficientNet-B0 with an accuracy of 0.86-0.88 over GoogleNet (~0.82-0.84), which also showed a lower convergence efficiency. While no significant difference was observed in the recall scores of EfficientNet between the weighted and unweighted loss methods, GoogleNet showed a considerable improvement in the true positive rates from 0.6177 to 0.7160 using the weighted loss strategy. Despite using the weighted loss method, there were negligible improvements in the performance of grid search-optimized networks. The poor classification results suggest low model generalization ability using manual tuning. This study concluded that the PSO-optimized cascaded segmentation-classification framework could potentially leverage advancements in human movement evaluation and rehabilitation assessment for sports applications. This system can be improved by adopting larger datasets or collaborative machine learning training to enhance model convergence for practical rehabilitation applications to assess users’ physical and fitness performance.

Keywords: Sports action, EfficientNet, Optimization, Segmentation, Grid search.




Audrey Huong (PhD)*, Faculty of Electrical and Electronic Engineering, Universiti Tun Hussein Onn Malaysia, 86400 Parit Raja, Batu Pahat Johor, Malaysia; Email: audrey@uthm.edu.my

Ser Lee Loh (PhD), Centre for Robotics and Industrial Automation, Fakulti Teknologi dan Kejuruteraan Elektrik, Universiti Teknikal Malaysia Melaka, Melaka, Malaysia; Email: slloh@utem.edu.my

Kok Beng Gan (PhD), Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, Bangi, Selangor, Malaysia; Email: kbgan@ukm.edu.my

Xavier Ngu (PhD), RF EMC Centre Malaysia Sdn. Bhd., Universiti Tun Hussein Onn Malaysia, 86400 Parit Raja, Batu Pahat, Johor, Malaysia; Email: xavier@uthm.edu.my

    [1] National Safety Council (2024). Sports and recreational injuries. Available: https://injuryfacts.nsc.org/home-and-community/safety-topics/sports-and-recreational-injuries (accessed on 3 Jan 2025).
    [2] Liao, Y.; Vakanski, A.; Xian, M.; Paul, D.; Baker, R. A review of computational approaches for evaluation of rehabilitation exercises. Comput. Biol. Med. 2020, 119.
    [3] Ebeling, P.R.; Cicuttini, F.; Scott, D.; Jones, G. Promoting mobility and healthy aging in men: a narrative review. Osteoporos Int. 2019, 30, 1911–1922.
    [4] Pan, C.T.; Lee, M.C.; Huang, J.S.; Chang, C.C.; Hoe, Z.Y.; Li, K.M. Active assistive design and multiaxis self-tuning control of a novel lower limb rehabilitation exoskeleton. Machines 2022, 10, 318.
    [5] Seçkin, A.Ç.; Ateş, B.; Seçkin, M. Review on wearable technology in sports: concepts, challenges and opportunities. Appl. Sci. 2023, 13, 10399.
    [6] Sierotowicz, M.; Connan, M.; Castellini, C. Human-in-the-loop assessment of an ultralight, low-cost body posture tracking device. Sensors 2020, 20, 890.
    [7] Yang, Y.; Meng, L. Physical education motion correction system based on virtual reality technology. Int. J. Emerg. Technol. Learn. 2019, 14, 105-116.
    [8] Li, C.H.J.; Liang, V.; Chow, Y.T.H.; Ng, H.Y.; Li, S.P. A mixed reality-based platform towards human-cyber-physical systems with iot wearable device for occupational safety and health training. Appl. Sci. 2022, 12, 12009.
    [9] Tavakoli, M.; Carriere, J.; Torabi, A. Smart wearable technologies, and autonomous intelligent systems for healthcare during the COVID-19 pandemic: An analysis of the state of the art and future vision. Adv. Intell. Syst. 2020, 2, 2000071.
    [10] Qiu, S.; Liu, L.; Zhao, H.; Wang, Z.; Jiang, Y. MEMS inertial sensors-based gait analysis for rehabilitation assessment via multi-sensor fusion. Micromachine 2018, 9, 442.
    [11] He, Z.; Liu, T.; Yi, J. A wearable sensing and training system: towards gait rehabilitation for elderly patients with knee osteoarthritis. in IEEE Sens. J. 2019, 19, 5936-5945.
    [12] Wang, H. Recognition of wrong sports movements based on deep neural network. IIETA 2020, 34, 663-671.
    [13] Zhang, L. Applying Deep Learning-Based Human Motion Recognition System in Sports Competition. Front. Neurorobot. 2022, 16, 1-12.
    [14] Hussain, A.; Khan, N.; Munsif, M.; Kim, M. J.; Baik, S. W. Medium Scale benchmark for cricket excited actions understanding. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, WA, USA, 2024.
    [15] Mottaghi, A.; Soryani, M.; Seifi, H. Action recognition in freestyle wrestling using silhouette-skeleton features. Eng. Sci. Technol. Int J. 2020, 23, 921-930.
    [16] Xiang, L.; Gao, X. Perceptual feature integration for sports dancing action scenery detection and optimization. in IEEE Access 2024, 12, 122101-122113.
    [17] Alavigharahbagh, A.; Hajihashemi, V.; Machado, J.J.M.; Tavares, J.M.R.S. Deep learning approach for human action recognition using a time saliency map based on motion features considering camera movement and shot in video image sequences. Information 2023, 14, 616.
    [18] Yuan, Y.; Qin, W.; Ibragimov, B.; Zhang, G.; Han, B.; Meng, M.Q.H.; Xing, L. Densely connected neural network with unbalanced discriminant and category sensitive constraints for polyp recognition. IEEE Trans. Autom. Sci. Eng. 2020, 17, 574–583.
    [19] Prayitno, C.R.; Shyu, K.T.; Putra, H.C.; Chen, Y.Y.; Tsai, K.S.; Hossain, M.T.; Jiang, W.; Shae Z.Y. A systematic review of federated learning in the healthcare area: From the perspective of data properties and applications. Appl. Sci. 2021, 11, 11191.
    [20] Liu, C.; Zhang, W.; Xu, W.; Lu, B.; Li, W.; Zhao, X. Substation Inspection safety risk identification based on synthetic data and spatiotemporal action detection. Sensors 2025, 25, 2720.
    [21] Alshirbaji, A.T; Jalal, N.A.; Docherty, P.D.; Neumuth, T.; Möller, K. Robustness of convolutional neural networks for surgical tool classification in laparoscopic videos from multiple sources and of multiple types: a systematic evaluation. Electronics 2022, 11, 2849.
    [22] Tsai, T.H.; Wang, C.Y. Wafer map defect classification using deep learning framework with data augmentation on imbalance datasets. J Image Video Proc. 2025, 6, 2025.
    [23] Kunal, K.; Tafeer, A.; Shantanu, G.; Rajat, G. Optimizing CNN hyperparameters for satellite image segmentation: A grid search approach. Int. J. Eng. Res. Manag. 2020, 4, 140–148.
    [24] Zhao, J.; Wang, J.; Han, M.; Fan, J. Adaptive particle swarm optimization of u-net parameters for marine aquaculture image segmentation. In 2025 13th International Conference on Intelligent Control and Information Processing (ICICIP), Muscat, Oman, 2025.
    [25] Zhang, W.; Liu, Z.; Zhou, L.; Leung, H.; Chan, A.B. Martial arts, dancing and sports dataset: A challenging stereo and multi-view dataset for 3D human pose estimation. Image Vis. Comput. 2017, 61, 22-39.
    [26] Yamany, W.; Moustafa N.; Turnbull, B. OQFL: An optimized quantum-based federated learning framework for defending against adversarial attacks in intelligent transportation systems. IEEE Trans. Intell. Transp. Syst. 2023, 24, 893-903.
    [27] Engelbrecht, A. P. In Computational Intelligence: An Introduction, 2nd ed.; Wiley, New Jersey, U.S., 2007.
    [28] Kuang, S.; Wang, L. Identification and analysis of consensus RNA motifs binding to the genome regulator CTCF. NAR Genom Bioinform. 2020, 6.
    [29] Huong, A.; Ngu, X. An optimized semantic segmentation framework for human skin detection. Int. J. Integr. Eng. 2024, 16, 293-300.

Audrey Huong (PhD), Ser Lee Loh (PhD), Kok Beng Gan (PhD), and Xavier Ngu (PhD) (2025), Cascaded Segmentation-Classification Framework Optimized for Sports Action Recognition Using Still Images: A Comparative Analysis of PSO and Grid Search. IJEER 13(4), 735-748. DOI: 10.37391/IJEER.130414.