| 摘要 |
A method for decoding an encoded image signal supplied in the form of a series of encoded image frames, each of the encoded image frames being divided into a multiplicity of macroblocks of pixels, each of the macroblocks having an associated half-pixel resolution motion vector which represents a translatory motion of the macroblock between a present and its preceding frames of the encoded image signal to a half-pixel accuracy, comprises: providing differential pixels of a differential macroblock in a first predetermined scanning order by performing entropy decoding, inverse quantization and inverse transformation on the pixels in each macroblock of the present frame; scanning the differential pixels so as to provide scan converted differential pixels of a scan converted differential macroblock in a second predetermined scanning order; providing half-pixel resolution pixels of a half-pixel resolution macroblock in the second predetermined scanning order by accessing, under a control of the half-pixel resolution motion vector, the pixels in the previous frame in an order corresponding to the second predetermined scanning order and, depending on the half-pixel resolution motion vector, performing a spatial interpolation on the pixels in the previous frame; and adding, on a pixel-by-pixel basis, the scan converted differential macroblock to the half-pixel resolution macroblock, to thereby form a reconstructed macroblock of the present frame.
|
