Music synchronization with video using emotion similarity
Ki-Ho Shin, In‐Kwon Lee · 2017
In this paper, we introduce a synchronization method of music and video, where the music is arranged with the video using emotion similarity. The music segments are matched with the video segments so that the music best follows the video's narrative. The temporal order of video segments are unchanged. Due to the disparity of structural musical content between segment boundaries, we design a cost function to overcome the problem. We perform a user study where the participants rate the overall quality of each synchronized video and show that our method generates a promising result.