In synthesizing live-action video and CG, when live-action video is shot by a moving camera, generation of CG that moves in accordance with the move of the live-action video without causing visual discomfort required a difficult operation such as programming in advance to precisely adjust the display timing and the display position of the CG. The present invention is to generate CG in accordance with live-action video shot by a moving camera with minimum visual discomfort, thereby creating a highly realistic synthesized video image.