School of Physical Education Institute (Main Campus), Zhengzhou University, No. 100 Science Avenue, Zhengzhou 450001, China.
Department of Physical Education, Sangmyung University, Seoul 390-711, Korea.
Int J Environ Res Public Health. 2021 Aug 27;18(17):9049. doi: 10.3390/ijerph18179049.
The recent curriculum reform in China puts forward higher requirements for the development of physical education. In order to further improve students' physical quality and motor skills, the traditional model was improved to address the lack of accuracy in motion recognition and detection of physical condition so as to assist teachers to improve students' physical quality. First, the physical education teaching activities required by the new curriculum reform were studied with regard to the actual needs of China's current social, political, and economic development; next, the application of artificial intelligence technology to physical education teaching activities was proposed; and finally, deep learning technology was studied and a human movement recognition model based on a long short-term memory (LSTM) neural network was established to identify the movement state of students in physical education teaching activities. The designed model includes three components: data acquisition, data calculation, and data visualization. The functions of each layer were introduced; then, the intelligent wearable system was adopted to detect the status of students and a feedback system was established to assist teaching; and finally, the dataset was constructed to train and test the designed model. The experimental results demonstrate that the recognition accuracy and loss value of the training model meet the practical requirements; in the algorithm test, the motion recognition accuracy of the designed model for different subjects was greater than 97.5%. Compared with the traditional human motion recognition algorithm, the designed model had a better recognition effect. Hence, the designed model can meet the actual needs of physical education. This exploration provides a new perspective for promoting the intelligent development of physical education.
中国最近的课程改革对体育教育的发展提出了更高的要求。为了进一步提高学生的身体素质和运动技能，改进了传统模式，以解决运动识别准确性和身体状况检测方面的不足，从而协助教师提高学生的身体素质。首先，研究了新课程改革所要求的体育教学活动，以满足中国当前社会、政治和经济发展的实际需求；其次，提出了将人工智能技术应用于体育教学活动；最后，研究了深度学习技术，并建立了基于长短时记忆（LSTM）神经网络的人体运动识别模型，以识别体育教学活动中学生的运动状态。所设计的模型包括三个部分：数据采集、数据计算和数据可视化。介绍了各层的功能；然后，采用智能可穿戴系统来检测学生的状态，并建立反馈系统以协助教学；最后，构建数据集来训练和测试所设计的模型。实验结果表明，训练模型的识别准确率和损失值符合实际要求；在算法测试中，所设计模型对不同主体的运动识别准确率均大于 97.5%。与传统的人体运动识别算法相比，所设计的模型具有更好的识别效果。因此，所设计的模型能够满足体育教育的实际需求。这一探索为推动体育教育的智能化发展提供了新视角。
