Author(s)

Huijian Deng ¹, Zijing Zhang ², Zheng Luo ³

Affiliation(s)

¹ Department of Physical Education, Chongqing University of Technology, Chongqing, 400054, China
² Department of Physical Education, Chongqing University of Science and Technology, Chongqing, 401331, China
³ Hunan Institute of Science and Technology, Yueyang, Hunan, 414006, China

Corresponding Author

Zheng Luo

ABSTRACT

This study aimed to develop an intelligent assessment method for evaluating body posture and movement timing during the stationary straight thrust in épée and to examine its performance, criterion validity, and instructional application. Fifteen experts participated in two rounds of Delphi consultation to establish the assessment framework. The model was developed using 1,200 videos collected from 60 students and was independently validated using 120 videos from an additional 20 students. To evaluate the instructional application of the model, 56 students were stratified by sex and pretest score and then randomly assigned to either a conventional-instruction group or a group receiving conventional instruction supplemented with intelligent feedback for eight weeks. The proposed model integrated single-person COCO-17 keypoint extraction, phase feature encoding, confidence gating, and multitask regression for posture and movement-timing assessment. The FPT+CPG model achieved a phase-recognition accuracy of 83.03%, a macro-averaged F1 score of 82.43%, a scoring mean absolute error (MAE) of 3.85 points, and a coefficient of determination (R²) of 0.845. The model also demonstrated strong agreement with the mean total scores assigned by experts, with an intraclass correlation coefficient (ICC(2,1)) of 0.925 and an MAE of 2.80 points. After adjustment for pretest scores, students who received intelligent feedback scored 6.08 points higher at posttest than those who received conventional instruction alone (95% confidence interval, 3.98–8.17; p < 0.001). The intelligent-feedback approach also reduced the mean feedback time to 14.49 ± 3.02 s, compared with 46.12 ± 6.01 s for conventional instruction (p < 0.001). These findings indicate that the joint modeling of movement-phase information and keypoint confidence can support the assessment of body posture and movement timing during the stationary straight thrust in épée. The results further suggest that instructor-reviewed intelligent feedback may improve movement performance and reduce feedback time when used as an adjunct to formative assessment in university fencing instruction.

KEYWORDS

Fencing instruction; Movement technique assessment; Human pose estimation; Movement timing; Formative feedback

CITE THIS PAPER

Huijian Deng, Zijing Zhang, Zheng Luo. Intelligent Assessment of Body Posture and Movement Timing in University Fencing Instruction: Development and Validation of a Phase-Aware, Confidence-Gated Model. Journal of Human Movement Science (2026) Vol. 7, No. 1, 80-94. DOI: http://dx.doi.org/10.23977/jhms.2026.070111.

REFERENCES

[1] Sun D, Song Y, Cen X, et al. Research progress of computer vision-based markerless sports motion capture technology. Journal of Shanghai University of Sport. 2021;45(9):70–85. doi:10.16099/j.sus.2021.09.007. [In Chinese].
[2] Yu L, Du C, Yan Z, et al. Review of 2D human pose encoding and decoding methods: from the perspective of ambiguity mitigation. Journal of Image and Graphics. 2024;29(11):3319–3344. doi:10.11834/jig.230648. [In Chinese].
[3] Huang Q, Cui J, Li C. A review of skeleton-based human action recognition. Journal of Computer-Aided Design & Computer Graphics. 2024;36(2):173–194. doi:10.3724/SP.J.1089.2024.2023-00358. [In Chinese].
[4] Cust EE, Sweeting AJ, Ball K, et al. Machine and deep learning for sport-specific movement recognition: a systematic review of model development and performance. Journal of Sports Sciences. 2019;37(5):568–600. doi:10.1080/02640414.2018.1521769.
[5] Malawski F, Kwolek B. Recognition of action dynamics in fencing using multimodal cues. Image and Vision Computing. 2018;75:1–10. doi:10.1016/j.imavis.2018.04.005.
[6] Malawski F, Krupa M, Kapela K. Automatic assessment of skill and performance in fencing footwork. In: Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications: VISAPP. Setúbal: SCITEPRESS; 2024. p. 637–644. doi:10.5220/0012577600003660.
[7] Lai C, Mo J, Xia H, et al. FACTS: fine-grained action classification for tactical sports. arXiv [Preprint]. 2024 Dec 21 [cited 2026 Jul 16]. arXiv:2412.16454. doi:10.48550/arXiv.2412.16454.
[8] Liu C, Zang Y, Lan C. Research on multiplayer pose estimation in complex sports scenes. Journal of University of Electronic Science and Technology of China. 2024;53(6):930–939. doi:10.12178/1001-0548.2023233. [In Chinese].
[9] Yan S, Xiong Y, Lin D. Spatial temporal graph convolutional networks for skeleton-based action recognition. Proceedings of the AAAI Conference on Artificial Intelligence. 2018;32(1):7444–7452. doi:10.1609/aaai.v32i1.12328.
[10] Chen Y, Zhang Z, Yuan C, et al. Channel-wise topology refinement graph convolution for skeleton-based action recognition. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. 2021. p. 13359–13368. doi:10.1109/ICCV48922.2021.01311.
[11] Duan H, Zhao Y, Chen K, et al. Revisiting skeleton-based action recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022. p. 2969–2978. doi:10.1109/CVPR52688.2022.00298.
[12] LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015;521(7553):436–444. doi:10.1038/nature14539.
[13] Fitts PM, Posner MI. Human performance. Belmont, CA: Brooks/Cole; 1967.
[14] Hattie J, Timperley H. The power of feedback. Review of Educational Research. 2007;77(1):81–112. doi:10.3102/003465430298487.
[15] Mishra P, Koehler MJ. Technological pedagogical content knowledge: a framework for teacher knowledge. Teachers College Record. 2006;108(6):1017–1054. doi:10.1111/j.1467-9620.2006.00684.x.
[16] Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine. 2016;15(2):155–163. doi:10.1016/j.jcm.2016.02.012.
[17] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307–310. doi:10.1016/S0140-6736(86)90837-8.
[18] Vickers AJ, Altman DG. Analysing controlled trials with baseline and follow up measurements. BMJ. 2001;323(7321):1123–1124. doi:10.1136/bmj.323.7321.1123.

Sponsors, Associates, and Links

---

• Journal of Language Testing & Assessment
  
  
• Information and Knowledge Management
  
  
• Military and Armament Science
  
  
• Media and Communication Research
  
  
• Art and Performance Letters
  
  
• Lecture Notes on History
  
  
• Lecture Notes on Language and Literature
  
  
• Philosophy Journal
  
  
• Science of Law Journal
  
  
• Journal of Political Science Research
  
  
• Journal of Sociology and Ethnology
  
  
• Advances in Broadcasting