Chart of Hitachi's technical process to super resolution
Chart of Hitachi's technical process to super resolution
[Click to enlarge image]

Hitachi, Ltd. has demonstrated its image signal processing technologies to achieve higher resolution view and smoother motion display. Both technologies are currently being developed in view of their introduction to future TVs.

The technology that boosts resolution view was demonstrated as "super resolution" high quality image processing technology. Hitachi demonstrated upgrading of 640 x 360 pixel progressive scanning video to 1280 x 720 pixels. The technology calculates a motion vector from images of two frames next to each other, and adjusts image location based on the calculation. Accordingly, the technology acquires twice many sample points and image data, compared to today's technology. Then the technology figures out image data to be actually displayed based on these data. As a result, Hitachi succeeded in displaying images at a nearly twice higher resolution than that achieved by the current technology.

As issues to tackle, Hitachi sited measures for video with little action as well as video with a lot of action. Although using a video that shows relatively smooth action in this demonstration, the company intends to increase the motion vector sensing precision for video that shows a lot of action in the near future.

As for the smooth motion display, the technology is geared for image signals derived from sources shot at 24 frames per second (fps) like movies. Generally, a method called "3:2 pulldown" is used to process such video sources and the same frame is repeatedly shown two or three times after the processing. Hitachi's new development firstly calculates original 24-fps image signals from images processed in the 3:2 pulldown mode. Then the technology figures out a motion vector of two frames next to each other and creates intermediate image signals between these two frames.