Research On Video Summarization Method Based On Submodular Function Optimization

With the development of network and multimedia,more and more video data are shared online on the Internet social platform.At the same time,video capture becomes very simple and cheap,and users follow a capture first,filter later mentality.Therefore,the videos obtained is long,unstable,redundant and unrepresentative,This paper mainly focuses on the research of dynamic video summarization algorithm,and proposes a new video summarization method to deal with the original and randomly captured video.The algorithm can input the original video,extract the useful information in the video,and output a short video summarization that can meet people’s needs but convey the story of the original video,The main work of this paper is as follows:(1)In this paper,the original diversity parameter is analyzed,and it is found that this parameter can not accurately represent the diversity of the video,resulting in a highly similarity of the generated summary content.On this basis,a new diversity parameter is proposed to reduce the similarity of the generated video summary segments.(2)It is proved that the new proposed diversity function is submodular.The proof of submodularity of three parameters of video summarization is also given.(3)Based on the original algorithm framework,a new video summarization optimization algorithm based on submodular function is designed by using the characteristics of submodular function.It can find the approximate solution of real summarization.(4)The experimental results show the efficiency of new proposed diversity parameter.We show that the similarity of the summarization given by the new method is better than that of the old methods.