Hierarchical video coding scheme with scalability and compatibility

In this paper, functional requirements for the MPEG 2 coding method are presented and, as an implementation method of these requirements, a hierarchical video coding scheme is proposed. In the proposed coding method, the prediction efficiency is increased by introducing spatial prediction in addition to temporal prediction as used in the MPEG 1 coding method. The hierarchical scheme is also characterized by the following features. (1) The input format is an output device‐independent noninterlaced video; (2) the coder has a hierarchical structure that permits the resolution level to be selected by the user; (3) by means of a small number of operations and a wide compensation area, the coder gives a motion‐detection method which detects motion vectors of each layer and a coding method of motion vectors; and (4) the proposed coding system introduces the use of down/up sampling processes with no phase distortion. Furthermore, the coding data description method is realized as an extension of that used by MPEG 1 and thus retains the compatibility. Simulation experiments show that this coding method satisfies the functional requirements without any obvious decrease in coding efficiency.