Research on Intelligent Recognition Method of Music Similar Segments Based on Deep Reinforcement Learning. Issue 3 (August 2021)

• Record Type: Journal Article
• Authors: Lu, Ni
• Is Part Of: Journal of physics. Volume 1992:Issue 3(2021)
• Publication Date: 2021-08
• Subjects: In-depth Reinforcement Learning -- Similar Segments -- Midi Music Files -- Pattern Recognition; Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1992/3/032041
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: The identification of similar music segments is of great significance for the study of online music search, content relevance, emotional expression and many other aspects. In the overall structure of music, the extraction of key frames, the identification of similar key frames for different types of music, which is to obtain better music emotion data. This paper uses in-depth reinforcement learning algorithms to analyze the music data in detail to construct music similarity Intelligently identify the database and match the obtained music files with the music data in the database to find similar segments. Case analysis shows that this method can effectively analyze music fragments and provide a basis for subsequent music control.