A systematic study of physical fitness assistance training for adolescents based on Kinect motion capture

Kinect motion capture technology records body motions, allowing for accurate monitoring and analysis in a variety of fields. This study investigates the intelligent recognition of classroom teaching behaviours by physical fitness instructors through the combination of Kinect sensors and machine learning algorithms. We proposed a novel Crayfish Optimization-driven Adaptive-Weighted AdaBoost (CO-AWAdaBoost) approach for classifying physical fitness instructional behaviours based on body posture data recorded by Kinect sensors. Z-score normalization is utilized to pre-process the obtained raw data. In our proposed recognition model, the CO algorithm leverages the natural behaviours of crayfish to optimize the process of feature selection. AdaBoost iteratively trains weak classifiers, assigning higher weights to misclassified samples. Our model can assist with the quantitative assessment of physical fitness classroom instruction, instructive suggestions, and large-scale behavioural investigation. The proposed detection model has been implemented in a Python program. In the results assessment phase; we evaluate our proposed model’s effectiveness in classifying physical fitness instructional behaviours using numerous evaluation metrics such as recall, F1-score, precision, and accuracy. During the finding evaluation phase, we thoroughly scrutinize the recognition effectiveness of the suggested model across various parameters, including precision (97.22%), accuracy (98.25%), specificity (97.85%), recall (97.86%), and F1-score (97.88%). We also carried out a comparison analysis with other traditional approaches. Our experimental findings demonstrate the reliability of the recommended framework.