영상구성 파라미터 기반 지능형 OST 매칭 프로그램 개발 연구

초록

In this study, an ‘intelligent OST matching program’ was developed to find background music using image recognition methods. To make this possible, convergence content was created by selecting scene analysis elements as ‘emotion,’ ‘color,’ and ‘pixel motion speed,’ and music components as ‘pitch,’ ‘tempo,’ and ‘dynamics.’ In the program, a ‘BGM library’ was created in advance and various music was classified. In addition, it was possible to select music that matches as much as possible by analyzing the input video and comparing the components with the analysis results of the music stored in the ‘BGM library’. In this process, a radial pentagonal graph was used to make it easier to see the comparison process between video and background music. Finally, it was possible to select ‘BGM Choice Sensitivity’ and ‘The Ratio of Background Music to Video Sound,’ and if a matching background music could not be found, music selected in advance could be used. I hope that the program produced in this way can be appropriately used for the creative activities of individual video producers.

키워드