Four-dimensional(4D) matrix and four-dimensional matrix DCT(4D-MDCT) theory are proposed and applied in color video coding

in order to obtain high compression for color video under high signal-to-noise ratio. The four-dimensional matrix (4D Matrix) theory is a new and useful theory for color video processing. The color video coding method based on 4D-MDCT put several frames of color video into a unified mathematical model (i.e.

4D matrix)

and various correlations in color video are eliminated utilizing 4D-MDCT. The four dimensions of 4D-MDCT used in the proposed algorithm are image width (row component)

image height (column component)

three channels (Y

U and V) in a color image (vertical component)

and adjacent three frames of the video sequence (depth component) respectively. Similar to the method of 2D-DCT

the whole three frames of a color video sequence are also divided into“4D-blocks”

that is 4D sub-matrices

and the order of each sub-matrix is 3×3×3×3. The transform coefficients are coded by vector quantization further. The method can fully exploit correlations between adjacent pixels of an image

YUV frames of color space

and adjacent frames of video at the same time. Experimental results has proved that for the color video in applications such as videophone and video conferencing

the proposed method can achieve higher compression ratio under high signal-to-noise ratio.