FIELD: information technology.
SUBSTANCE: invention relates to video encoding and decoding. Image decoding method comprises steps of: determining whether reference picture and a candidate block, and reference picture of current long-term prediction reference unit images, candidate block is one of a plurality of candidate blocks adjacent to current prediction unit; and when reference picture candidate block and current reference picture prediction unit are long-term reference pictures, is obtained for candidate spatial prediction motion vector, not by scaling motion vector candidate block; determining current vector prediction motion prediction unit from among candidates for prediction of motion vector candidate having spatial prediction motion vector, and generating motion vector prediction of current unit by using a prediction motion vector.
EFFECT: technical result is high efficiency of motion vector prediction processing.
3 cl, 28 dwg, 1 tbl
