Traditional assessment methods rely on teachers’ subjective judgment and lack objective quantitative standards. With the popularization of piano education, scientific assessment of students’ piano skill level has become particularly important. This study proposes an automatic piano skill assessment framework based on computer vision motion tracking and ID3 decision tree classification model to capture the motion characteristics of piano playing and analyze them quantitatively. The research methodology captured feature data such as keystroke and tone control, pedal technique, and musical expression through visual recognition assessment of 42 piano performers of different levels, and established an ID3 decision tree classification model to automatically assess students’ piano skill level. The results showed that: the accuracy of the visual recognition-based assessment reached 81.7%, with a Wilks’ Lambda value of 0.25; the ID3 decision tree classification model reached 95.8% accuracy on the training set and 74.5% on the test set; it was found by Pearson’s correlation coefficient analysis that keystroke and tone control (0.411) and pedal technique (0.408) were the most important features for discriminating the the most important characteristic parameters of piano skill level. Upon comparison with other methods, the assessment framework proposed in this study outperformed the existing methods in terms of accuracy, with an assessment accuracy of 84.86%. The study shows that the pattern recognition technology combining computer vision and ID3 decision tree can effectively realize the objective assessment of piano skill level and provide a scientific and quantitative evaluation system for piano teaching.
