| With the development of three-dimensional display technology and multimedia technology, many manufacturers have launched their own three-dimensional display, three-dimensional video technology have become an increasing concern.3D-TV, 3D-film and 3D advertising is no longer strange for people, and it has achieved a huge interest. The earlier 3DTV needed the equipment (such as 3D glasses, helmets, etc.) to watch the 3D effects. Now it has emerged as the device that does not need the help of any aid you can watch the 3D effects. Currently, there are two restricting factors of popularity of 3DTV, one is the system program of 3DTV, the other is the content of 3D video, and people pay more attention to the conversion of 2D video to 3D video.This paper describes the principle of 3DTV, the development of the 3DTV system and a natural 3DTV system. From the system of 3DTV, we can see that the system based on depth information may become a standard of 3DTV in the future, so this paper proposed a method of semi-automatic video conversion of 2D to 3D based on depth information after introducing two methods of video conversion of 2D to 3D.The method of semi-automatic conversion divided video sequences into key frames and non-key frames. We get the depth information of key frames artificially, and get the depth information of non-key frames automatically. We study the common block matching motion estimation algorithm and combine with the method of SEA, propose a method of SEA-DS, this method reduces the average number of searches under the premise of guaranteeing quality, which is an improve of motion estimation algorithm.Get the motion vectors using SEA-DS in the method of semi-automatic conversion, then generate the depth information of non-key frame and its feasibility is verified by experiments, which solves the problem of generating depth information of non-key automatically in a small area. This method improves the conversion speed and reduces the production workload of 3D video content. |
PDF Full Text Request