| 摘要 |
A frame of a video signal, wherein the frame is divided into non-overlapping blocks of MxN pixels, M and N being positive integers, respectively, is encoded by: arranging the blocks into groups of PxQ blocks, P and Q being positive integers, respectively; transforming each of the PxQ blocks in a group into a set of MxN transform coefficients to thereby provide for the group a first class of transform coefficients corresponding to PxQ sets of MxN transform coefficients, transform coefficients in a set representing different frequency components, respectively; rearranging the first class of transform coefficients to generate MxN sets of PxQ rearranged coefficients, the rearranged coefficients in a set being of an identical frequency component; transforming each set in the MxN sets of PxQ rearranged coefficients into a set of PxQ transform coefficients to thereby provide a second class of transform coefficients corresponding to the transform coefficients of the MxN sets of PxQ rearranged coefficients; and setting the second class of transform coefficients as selected coefficients for the group of PxQ blocks.
|
