Video coder and decoder using adaptive lattice quantiser

摘要

The video coding and decoding is based on an adaptive lattice quantiser. An input image from a video sequence is divided into blocks of pixel data forming vectors. The set of vectors is divided into blocks, or macroblocks. The vectors in each macroblock are subject to a vectorial lattice quantisation which examines the range of vectorial components in all dimensions, and adjusts the density of the lattice vectors as a function of the range of the vectorial components. Each quantised vector in the macroblock is converted to an index of the lattice vector. Service information concerning the range in a macroblock, and a set of indexes, is transmitted to holding blocks and stored there.