Method for scenery model aided image data reduction for digital television signals

摘要

In this method an image signal supplied at time t is to be coded, a preceding image, already coded at time t-1, from a scene being present in a frame buffer as reference and the frame-to-frame information consisting of an amplification, a displacement and an adaptively obtained quad-tree division structure in which it is provided that when the system to be operated by the method is initialised, a uniformly predetermined grey value expressed as a particular luminance value is entered completely in the same manner, that is to say over all pixels, in the frame buffer of a coder at the transmitting end and in the frame buffer of a decoder at the receiving end and that both the frame buffers in the coder and the frame buffer in the decoder are in each case operated fed back to themselves in such a manner that the content of the frame buffers in the coder and decoder can be read out in blocks of variable size, amplified by a factor greater than or less than 1 of the luminance and, addressed in a displaced manner, written back into the frame buffers, the variable-size blocks being organised in accordance with a quad-tree data structure known per se. …IMAGE…