Generating a set of motion vector predictors where determination of temporal predictors is dependent on selected spatial predictors

摘要

A set of motion vector predictors for a block to encode or decode of a current frame, is generated. The set is permitted to include at least one spatial motion vector predictor and at least one temporal motion vector predictor. A temporal motion vector is selected from a bottom right position of a block of the reference frame collocated with said block to encode or decode, or alternatively is selected from a predetermined position in the vicinity of said bottom right position, in a block of the reference frame neighbouring the collocated block. At least one said spatial predictor is selected before the temporal predictor is determined, and the determination of the temporal predictor is carried out in dependence upon the selected spatial predictor(s). The availability of the spatial predictors(s) may be determined and the temporal predictors may be selected in dependence upon this availability. This motion vector predictor selection scheme is then preferably used to encode a sequence of digital images into a bitstream, for subsequent decoding. Coding efficiency improvements can thereby be achieved.