NHK Science & Technology Research Laboratories are conducting R&D on integral 3D TV, the next broadcast service to follow on super-high definition.
They introduced the newest technology to create integral 3D content from multi-perspective videos at Technology Open House 2011.
“Conventional integral photography (IP) uses a dedicated IP camera that combines a super-high definition camera and lens array to film. For this, we took a different approach. We created a 3D model from multi-perspective videos of the object taken by multiple cameras, and are displaying the 3D model as an integral 3D video.”
“This video is actually not of a CG character. It is of an actor wearing a costume and surrounded by multiple cameras in a studio and filmed to create a 3D model from multi-perspective videos. The 3D video was overlapped with a CG background of an open field as you can see.”
The resolution of this display is 160×118 pixels, and the array has the same number of lenses. The video is projected onto a 4Kx2K display from behind the lens array.
The Labs used 40 cameras to generate the 3D model. They developed a method to project parallel light beams on a projection surface for the conversion to integral autostereoscopic content, so it is possible to achieve efficient conversion in which most of the processing is handled in the GPU to achieve high speeds.
“Since we could achieve high speeds in processing integral image generation from a 3D model, we can create interactive content. There is an acceleration sensor at the top of the display, and we are creating content that has the characters slipping smoothly depending on the direction of that sensor.”
See the original story here: http://www.diginfo.tv/2011/06/09/11-0118-r-en.php