An adaptive hybrid DPCM/DCT method for video coding

摘要

Discrete cosine transform (DCT) is a very popular coding algorithm for high compression ratio bandwidth reduction of digital video image sequences. Basically, DCT converts statistically dependent picture elements (pixels) into independent coefficients. The results are usually energy concentrated in only a few of the coefficients containing the main part of the picture information which can be quantized and run-length encoded. This concept is, however, not easy to apply when the transform pixel block is small in size and contains an edge boundary of an object. Transformation of such an edge block will not lead to an efficient compaction of signal energy. In this paper, an adaptive hybrid DPCM/DCT coder for video compression is presented. Having avoided using DCT coding for the edge blocks, the adaptive coder which instead uses a DPCM method on these edge blocks has reduced the ringing effects (corona effects) at the edges usual to transform coding method.