<![CDATA[This study examined the effectiveness of an artificial-intelligence (AI) learning medium using pose-estimation technology to improve elementary school students’ fundamental movement skills in physical education. A quasi-experimental nonequivalent control-group design was employed. Participants were 60 fifth-grade students assigned to an experimental group (n=30) and a control group (n=30). For six weeks, the experimental group learned with an AI medium that detected body keypoints, compared joint angles and movement sequences with reference models, and provided immediate visual feedback. The control group received teacher demonstrations and conventional video instruction. Fundamental movement skills were assessed through an observation rubric covering locomotor and non-locomotor components. Data were analyzed using paired-samples t tests, an independent-samples t test, normalized gain, and Cohen’s d. The experimental group improved from 58.43±7.84 to 82.17±9.33, while the control group improved from 57.77±7.61 to 70.53±11.11. The difference in gain scores was statistically significant (t=5.679; p<.001), with a large effect size (d=1.47). The AI medium was particularly useful for correcting knee–hip alignment, trunk position, step rhythm, and balance stability. These findings indicate that pose estimation can serve as augmented feedback that supports more objective movement monitoring. Nevertheless, its pedagogical use requires teacher supervision, appropriate camera placement, adequate lighting, and robust student-data privacy procedures.]]>
Copyrights © 2024
