Abstract: A process for integrating video images includes generating a dancer video with a dancer and identifying a dancer head position of the dancer in the dancer video. A participant head location is identified and a cropped area is generated based on the participant head location. A participant video is generated of cropped area contents and a composite video is generated by overlaying the cropped area contents onto the dancer head position. A video image integration system includes a video camera for capturing participant videos and a processor. The processor has a processor database module storing pre-recorded dancer videos, a face recognition module configured to identify participant head locations from the video camera and generated cropped areas based thereon, and a blender module configured to overlay the cropped areas onto the dancer videos to generate composite videos.

Abstract: A centralised game server in a bank (50) of game servers runs a game program for use by a user at a remote terminal (52, 56, 58). In the game server, the game program sends a first set of graphics instructions to a first graphics processing unit (76) which is intercepted by an instruction interception module (74). The first set of instructions, including vertex data, transformation data and texture data are passed to the first graphics processing unit (76) whilst a specially manipulated version of the instructions is generated and passed to a second graphics processing unit (78). The first graphics processing unit (76) renders the image data as the game intended whilst the second graphics processing unit (78) is used to render specially adapted graphics data from which to extract compression assistance data used for compression, e.g. motion vectors.