摘要
对视频内容添加适当的背景音乐,通常需要丰富的专业知识。面向这一应用,提出了一种基于视频场景分析的背景音乐自动推荐方法。该方法从场景、节奏、旋律和音色上进行特征分析,通过建立结构化关联模型,实现推荐音乐与视频内容的最佳匹配。实验结果表明,该方法能够有效地为个人录制视频自动推荐专业背景音乐,从而大大减少视频用户编辑背景音乐的工作量。
Embedding suitable background music into personal videos requires much rich professional knowledge. To this purpose, this paper proposed an automatic background music recommendation method based on the video scene analysis. The proposed method targets to make the recommended background music match the video content well, in terms of rhythm, melody, and timbre. Through the experiments on real videos, it is well verified that the proposed method can effectively recommend professional background music for various videos, which thus can significantly relieve the editing burden of the common users.
出处
《计算机应用》
CSCD
北大核心
2014年第A01期268-269,326,共3页
journal of Computer Applications
关键词
背景音乐
推荐算法
视频场景分析
支持向量机
视频特征
background narsic
recommendation method
video scene analysis
Support Vector Machine （SVM）
video feature
