Media Vision

James M Rafferty

A frame of digitized video data is compressed by dividing the frame into a plurality of rectangular blocks each typically measuring 4 pixels by 4 pixels. The entropy of each block is determined by measuring the difference between a maximum luminance of any pixel in the block and the minimum luminance of any pixel in the block. Those blocks which have a luminance value above a predetermined threshold (such as the average luminance for the block) are then subdivided into four blocks each including four pixels. Then, for each sub-block of 4 pixels a conventional video color compression technique is applied. For the non-subdivided blocks i.e., where the maximum luminance difference is below the threshold, the entire block is video compressed into a high color value and a low color value which are typically averages of the higher luminance and lower luminance group pixels respectively. Thus, some blocks are adaptively subdivided or not with greater video compression being provided for those blocks with relatively smaller variations in internal video information between the pixels and with the other blocks being subdivided to provide more video information and hence less compression. This results in a relatively high video compression ratio especially suitable for implementation by computer software.

Growing community of inventors

Redwood City, CA, United States of America

James M Rafferty