Search

A Computer Vision-Enabled Closed-Loop Framework for Adaptive Practical-Skill Instruction in Higher Vocational Education: A Rail Transit Maintenance Case

Author(s)

Yingge Li*, Ruihao Guo, Hongmei Liu, Weixun Chen

Affiliation(s)

Department of Guangzhou Railway Polytechnic, Guangzhou, China *Corresponding Author

Abstract

Practical training is central to talent cultivation in vocational education, yet adaptive instruction for practical skills remains difficult to implement at scale. Taking rail transit programs in Chinese higher vocational colleges as an example, this framework study identifies three major challenges: large individual differences in skill acquisition, insufficient fine-grained perception of operational processes, and limited capacity for targeted remediation. To address the data gaps inherent in traditional learner profiling, which primarily relies on online logs and test results, this paper proposes an adaptive practical skills teaching application model based on computer vision. This model is constructed around a closed-loop structure of collection-diagnosis-remediation-assessment, encompassing four levels: data foundation, skill diagnosis, targeted remediation, and outcome assessment, while also incorporating a risk management mechanism. The model focuses on five core dimensions of practical-skill diagnosis and assessment: operational compliance, safety compliance, action quality, job competence, and skill internalization. By extending learner-state data from online behavior to visual and multimodal evidence generated during practical operations, the model makes the operational process easier to observe and diagnose. This paper uses a rail transit maintenance training scenario to illustrate the implementation path of the model. This model provides a practical framework for data-driven, personalized skill instruction in higher vocational education.

Keywords

Computer Vision; Practical Skills; Learning Diagnosis; Digital Transformation of Vocational Education

References

[1]Bao X. D. Urban Rail Transit Present Situation and Future Development Trends in China: Overall Analysis Based on National Policies and Strategic Plans in 2016-2020. Urban Rail Transit, 2018, 4: 1-12. [2]Lu K., Zhang L., Li S., Huang Y. P. Ding X, Hao J N, Huang S Q, Li X J, Lu F, Zhang H W. Urban Rail Transit in China: Progress Report and Analysis (2015-2023). Urban Rail Transit, 2025, 11: 1-27. [3]Sun B. F., Liu J. J., Hao J. Y., Shen X. X., Mao X. H., Song X. M. Maintenance Decision-Making of an Urban Rail Transit System in a Regionalized Network-Wide Perspective. Sustainability, 2020, 12(22): 9734. [4]Yu S. Q. The Key Path of Educational Digital Transformation. Journal of East China Normal University (Educational Sciences), 2023, 41(3): 62-71. [5]Zhu Z. T., Hu J. The Practical Logic and Development Opportunities of Educational Digital Transformation. e-Education Research, 2022, 43(1): 5-15. [6]Yang X. M., Zhang Y. How to Resolve the Contradiction Between Scale and Personalization in Education: The Logical Framework and Practical Path of Data-Driven Large-Scale Adaptive Instruction. Distance Education in China, 2022(8): 42-52+79. [7]Liu H. Y., Zhang T. C., Wu P. W., Yu G.. A Review of Knowledge Tracing. Journal of East China Normal University (Natural Science), 2019(5): 1-15. [8]Liu T. Y., Chen W., Chang L., Gu T. L. Research Advances in Knowledge Tracing Based on Deep Learning. Journal of Computer Research and Development, 2022, 59(1): 81-104. [9]Wang Y., Zhu M. X., Yang S. H., Lu X. S., Zhou A. Y. A Survey and Performance Comparison of Deep Knowledge Tracing Models. Journal of Software, 2023, 34(3): 1365-1395. [10]Yang W. Y., Yang Y. H. Focusing on Interpretability: A Review and Outlook of Knowledge Tracing Models. Modern Educational Technology, 2024, 34(5): 53-63. [11]Shen Y. Y., Zhang Y. M., Shen Y. F. A Survey of Vision-Based Human Action Quality Assessment. Acta Automatica Sinica, 2025, 51(2): 404-426. [12]Wu J. C., Wang L. M., Wu G. S. A Survey of Video Group Activity Recognition. Journal of Software, 2023, 34(2): 964-984. [13]Zhang Y., Wen G. Z., Mi S. Y., Zhang M. L., Geng X.. A Survey of 2D Human Pose Estimation Based on Deep Learning. Journal of Software, 2022, 33(11): 4173-4191. [14]Science and Technology Development Center, Ministry of Education. Construction Guideline for Demonstration Virtual Simulation Training Bases in Vocational Education. 2021. [15]Han X. B., Shi W. R. Analysis of the Dilemmas and Solutions of Virtual Simulation Training in Vocational Education. e-Education Research, 2025(9): 50-56. [16]Ericsson K. A., Krampe R. T. Tesch-Romer C. The role of deliberate practice in the acquisition of expert performance. Psychological Review,1993,100(3):363-406.