Video Highlight Based Optimization And Transmission Research

With the rapid development of mobile Internet,the increasing popularity of mobile smart devices,the explosive growth of video applications such as video broadcast and on-demand video streaming,the proportion of video traffic in the entire network traffic is increasing.In order to make users get good quality of experience during video watching and gain some commercial value,the research on video transmission and optimization has aroused widespread concern in industry and academia.Even though the wireless communication technology has been rapidly developed in recent years and the wireless communication equipment has been improved day by day,the capacity of the network is constantly increasing but it still cannot meet the consumption of the network bandwidth resources caused by the large-scale video data generated by a growing number of video applications.Bandwidth resources are still very tight.On the other hand,the value of video data density is low,the video segment that really interests the user or reflects the main content of the video only occupies a small part of the entire video.Thus for the user,this small part of the video clip has the most important impact on the quality of the user's experience.The current video transmission scheme does not differentiate transmission of different video segments in the video,but uses the same transmission strategy,which results in the already insufficient bandwidth resources exerting the greatest potential to transmit videos that are of most importance to user experience quality Fragment,making the quality of the user's experience decreased.This paper conducted a video optimization and transmission research based on video highlight.Firstly,a video highlight detection method based on category cleaning was proposed.The K-means method was used to reduce the influence of the positive sample data and improve the detection of the importance of the video segment.Then,a dynamic adaptive streaming scheme based on the importance of video segments is proposed.Based on Dynamic Adaptive Streaming over HTTP(DASH)streaming media protocol,an adaptive code rate adaptive selection strategy based on importance of video segments is designed.The video segment of high importance chooses the video version of the high bit rate to improve the user's quality of experience.The simulation results show that our video-based segment detection method based on category cleaning has a better performance than the traditional method of video segment detection,especially when the proportion of high-importance video segments is small.In addition,our dynamic adaptive streaming scheme based on the importance of the video segment has a better quality of experience for users in the high-importance video segment than the traditional scheme.