Toshiba developed "2D 3 D conversion technology" that generates natural and beautiful 3D video from 2D video, "CELL REGZA" for 3D



Despite the release of 3D television this year, the fact is that the 2D video content is far more overwhelming than it is today, and even if you use the technology to convert 2D video to 3D video, 3D video Although the stereoscopic effect was poor with respect to the produced content, it is possible to generate natural and beautiful 3D video from 2D video by using advanced calculation processing2D 3 D conversion technologyToshiba announced that it has developed.

In addition, "1 million yen REGZA" that appeared last year "CELL REGZA"The latest models of three models also appear.

Details are as below.
Toshiba: News Release (2010 - 07 - 28): On the development of "2D 3 D conversion technology" to generate natural and beautiful 3D video from 2D video

According to this release, Toshiba has developed a technology to generate natural and beautiful 3D video which makes it easier to obtain more stereoscopic effect by estimating the depth with high precision from multiple elements in 2D video

The technology developed this time applied proprietary technologies such as "Motion Detection Technology" and "Face Detection Technology" of the company, the majority of television broadcasting is "part accompanying movement" "part where a person appears" "typical composition The landscape is classified into three categories, "Landscape" and "Landscape", and by combining them by applying optimal depth estimation factors to each, it generates 3D images with more stereoscopic effect.

Depth generation based on motion detection. We estimate the anteroposterior relation of an object based on the basic principle that "the closer the object is to the object, the faster the apparent motion is, the larger the object the object moves on the screen, the closer it is to the object."


Next, depth generation of a person based on face detection. We detect the position of the face from the person in the 2D video, statistically convert the relative position of the body such as the face and shoulder to data and assign it to the template to generate the depth of the person.


Finally, depth generation based on composition estimation. We estimate composition by using contrast of color distribution of displayed images such as sky, sea, land etc. in natural landscape and generate depth.


It is like this when it is in flow. By estimating the depth from three technologies, natural 3D images are produced from 2D images.


The "2D 3 D conversion technology" developed this time is a technology realized by using the Cell processor adopted also by Sony's PS 3 etc. It is a 3D compatible liquid crystal TV which will be released sequentially from the beginning of October It is supposed to be installed in "CELL REGZA 55X2 / 55XE2 / 46XE2".

This is "CELL REGZA 55X2". Adopted "CELL REGZA 3D Super Resolution Technology" which carries "4 × speed mega LED panel" of high brightness liquid crystal with dynamic contrast of 9 million: 1 and displays 3D broadcast image as high definition 3D image.


It adopts "wide color gamut VA system full HD clear LED panel" with dynamic contrast of 4 million: 1, and slimmed "55 XE 2" and "46 X E 2" will also be released at the same time. Like "55 X 2", the recording function is supported such as "time shift machine 2" that enables simultaneous recording of up to 8 channels of terrestrial digital broadcasting on a 3 TB HDD built into the main unit and normal recording to USB external HDD It is substantial. By the way, this is "55XE2".


"46 XE 2"


Details of "CELL REGZA" are from the following.

Toshiba: News Release (2010 - 07 - 28): On the launch of "CELL REGZA" liquid crystal television adopting 3D super resolution technology to display 3D broadcasting in high definition

in Hardware, Posted by darkhorse_log