VIDEO DECODING METHOD Russian patent published in 2019 - IPC H04N19/51 H04N19/122 H04N19/593 

FIELD: physics.

SUBSTANCE: invention relates to encoding and decoding a video signal based on a merging candidate unit. Technical result is achieved by forming a candidate merge list for the current prediction unit using the merging candidate unit, wherein the merging candidate unit includes a spatial merging candidate for fusion and / or a time merging candidate unit, performing inter-frame prediction of the current prediction unit based on the candidate fusion candidate list, the candidate spatial merging unit including at least one of neighbouring units adjacent to current prediction unit, wherein neighbouring units include at least one of a number of a left neighbouring unit, an upper neighbouring unit, a top right neighbouring unit, a lower left neighbouring unit and the upper left neighbouring unit, the merging candidate time unit contained in the image having a different time order compared to the current image containing the current prediction unit, size of motion estimation region (MER) is 8 × 8, and coding unit size is 8 × 8, then at least one of adjacent units is replaced by replacement unit located outside MER.

EFFECT: technical result consists in improvement of accuracy of video signal processing.

5 cl, 8 dwg, 2 tbl