Variable Length Segmentation Technique for Dynamic Video Streaming

Abstract

To transfer videos using dynamic streaming approach, they must be divided into segments; these segments would be compressed to different quality levels to be delivered from the source to the destination, with fewer amounts of delay and packet loss according to the network’s statues at the sending time. The almost used segmentation techniques depend on the time that the segment time length is constant within the same video file. This paper presents a suggestion to take into consideration the contents of the video to split the video file into segments. The length of the segments depends on the length of video scenes, which made the scene length varies according to the content of the video file itself.