Video Background Music Recognition and Automatic Recommendation Based on Gmm Model

• Published in: Informatica (Ljubljana) Vol. 47; no. 7; pp. 41 - 49
• Main Authors: Zhou, Wei, Ma, Ke
• Format: Journal Article
• Language: English
• Published: Ljubljana Slovenian Society Informatika / Slovensko drustvo Informatika 27.07.2023
• Subjects: Audio data; Cultural heritage; Deep learning; Emotions; Frequency response; Internet; Listening; Music; Musical performances; Neural networks; Principal components analysis; Probabilistic models; Recognition; Recommender systems; Signal processing; Smartphones; Social networks; Streaming media; Video; China
• ISSN: 0350-5596; 1854-3871
• DOI: 10.31449/inf.v47i7.4812

Recognizing background music in videos is a widely utilized technology in the global music business. With the use of classification, the data about the audio signal's frequency response, orchestration, and temporal structure is represented. In the beginning, identification was a human process. This operation may now be carried out autonomously because of developments in technologies and signal-processing techniques. Due to the widespread utilization of social networks, many smartphones come with a video-shooting feature that people often employ to create user-generated entertainment and communicate it with others. Nonetheless, it might be difficult to choose background music that complements the subject. Those who want to include background music in their videos must actively search for the audio. Nevertheless, since it is a procedure that requires a lot of time and effort, the emphasis of this study is on the construction of a system that will assist people in more easily and quickly obtaining the proper background music for their interests. For automatic recognition of video background music and recommendation, we implemented a Gaussian mixture model (GMM). Using principal component analysis, audio characteristics were recovered for effective recognition. The outcomes were assessed using performance measures and contrasted with previously used methods. The findings indicate that the suggested GMMproduces superior performance.