Application Analysis of Music Video Retrieval Technology Based on Dynamic Programming in Piano Performance Teaching

With the development of Internet technology, music videos on the network are becoming increasingly rich. How to extract concert video clips for specific scenes or shots from massive video libraries or ultra-long video files is a relatively difficult issue. Traditional music video retrieval methods are mostly based on key text retrieval. However, they cannot meet the needs of users. At the same time, in response to the demand for specific videos in piano performance teaching, it is also difficult for these methods to filter out key music clips from numerous videos. Therefore, a music video retrieval technology is constructed based on video feature similarity calculation. Aiming at the shortcomings of video similarity calculation methods, a dynamic programming algorithm is used to improve it. The improved music video retrieval technology is applied to the classroom learning practice of piano performance teaching, verifying the actual effect of this technology. The experimental results show that the accuracy of the music video retrieval technology reaches 91.02%. After being applied to piano classroom teaching, the overall performance of students has been improved. This shows that the proposed music video retrieval technology can effectively achieve the retrieval of required videos and improve the effectiveness of piano classroom teaching.