Fast generation of video hologram patterns by use of motion vectors of three-dimensional objects

摘要

Thus far, various approaches to generate the computer-generated holograms (CGHs) of 3-D objects have been suggested but, most of them have been applied to the still images, not to the video images due to their computational complexity. Recently, a method to fast compute the CGH patterns of 3-D video images has been proposed by combined use of data compression and novel look-up table (N-LUT) techniques. In this method, temporally redundant data of 3-D video images are removed with the differential pulse code modulation (DPCM) algorithm and then the CGH patterns for these compressed video images are calculated with the N-LUT method. However, as the 3-D objects move rapidly, image differences between the video frames may increase, which results in a massive growth of calculation time of the video holograms. Therefore, we propose a novel approach to significantly reduce the computation time of 3-D video holograms by employing a new concept of motion-vector of the 3-D object. In the proposed method, 3-D objects are firstly segmented from the 1st frame of the 3-D videos, and the CGH patterns for each segmented object are computed with the N-LUT algorithm. Secondly, motion vectors between each segmented object and the corresponding objects in the consecutive 3-D video frames are calculated. Thirdly, the CGH patterns for each segmented object are shifted with the calculated motion vectors. Finally, all these shifted CGH patterns are added up to generate the hologram patterns of the consecutive 3-D video frames. To confirm the feasibility of the proposed method, experiments are performed and the results are comparatively discussed with the conventional methods in terms of the number of object points and computation time.