Method and device for carrying out viewpoint prediction and realizing optimized transmission of virtual reality video by using cross user behavior mode

The invention provides a method and device for carrying out viewpoint prediction and realizing optimized transmission of virtual reality video by using a cross user behavior mode. Processing includingtime slicing and the spatial blocking and transcoding is performed on a panoramic video at a server end and appropriate video clips are selected in an adaptive manner at a client end for requesting.An adaptive transmission process includes two aspects: corresponding video content is continuously transmitted according to a user observation direction in an adaptive manner so as to realize predication on viewpoints; and the total code rate of the video is adjusted continuously in an adaptive manner so as to meet change of current network bandwidth. In order to properly implement the two adaptive processes, the invention provides the novel method and device for carrying out viewpoint prediction and realizing optimized transmission of virtual reality video by using the cross user behavior mode. The video quality of a