Innovative Practice of Physical Education Teaching in Colleges and Universities Based on Artificial Intelligence Technology

With the development and improvement of artificial intelligence technology, the teaching innovation mode of combining college sports courses with artificial intelligence has gradually received widespread attention. This paper is based on artificial intelligence technology for the design of intelligent sports detection wearable devices, through which students’ sports data are collected, low-pass filters are used to reduce the noise of the collected data, normalization is carried out, and the AlphaPose algorithm is combined with the assessment and extraction of the human movement posture of college sports. The artificial intelligence sports teaching framework is built to innovate college sports teaching based on artificial intelligence. Finally, the impact of sports recognition is examined, and a comparison experiment is carried out to examine the practical implications of this teaching method. The four experimental targets had sports recognition errors that were less than 10% on average according to the results. The two classes under the experimental control have a P-value of less than 0.05 in the comparison of physical skills and physical fitness test data, and there is an improvement of 5-20 in all scores, which indicates that the experimental class has a higher teaching effect and is significantly helpful in performance improvement. Based on the above, this paper researches the practice of artificial intelligence technology in college sports teaching to provide an innovative path for the transformation of traditional sports teaching to artificial intelligence sports teaching.