Motion vector prediction method and apparatus for encoding or decoding video

Abstract

Provided are a motion vector predicting method and an apparatus for encoding and decoding a video. The motion vector prediction method includes: determining, from neighboring blocks of the current block, a plurality of candidate blocks that are referred to so as to predict a motion vector of a current block; determining a candidate motion vector of a first candidate block among the determined plurality of candidate blocks, based on whether a reference image of the first candidate block and a reference image of the current block are long-term reference images; and determining the motion vector of the current block by using a candidate motion vector list including the determined candidate motion vector of the first candidate block and candidate motion vectors from remaining candidate blocks among the determined plurality of candidate blocks.

1 . An apparatus for decoding an image, the apparatus comprising:
 at least one processor configure to:   determine a collocated picture from among pictures restored before a current picture according to a collocated index;   parse information about a long-term reference picture from a bitstream;   when one of a reference picture of a collocated block included in the collocated picture and a reference picture of a current block is determined as a long-term reference picture, and the other one of the reference picture of the collocated block and the reference picture of the current block is determined as a short-term reference picture based on the information about the long-term reference picture, determine that a motion vector of the collocated block is un-available so that the motion vector of the collocated block is not used to predict a motion vector of the current block;   when both of the reference picture of the collocated block and the reference picture of the current block are determined as long-term reference pictures based on the information about the long-term reference picture, obtain a temporal motion vector prediction candidate without scaling the motion vector of the collocated block;   receive prediction information of a candidate block as prediction information of the current block that indicates the candidate block used to derive a motion vector predictor of the current block;   determine the motion vector predictor of the current block from among motion vector prediction candidates, comprising the temporal motion vector prediction candidate, based on the prediction information of the candidate block; and   generate the motion vector of the current block using the motion vector predictor,   wherein the scaling is based on a ratio of a distance (Td) between the collocated picture and the reference picture of the collocated block and a distance (Tb) between the current picture and the reference picture of the current block.   
     
     
         2 . An apparatus for encoding an image, the apparatus comprising:
 at least one processor configure to:   generate a collocated index for determining a collocated picture from among pictures restored before a current picture;   when one of a reference picture of a collocated block included in the collocated picture and a reference picture of a current block is determined as a long-term reference picture, and the other one of the reference picture of the collocated block and the reference picture of the current block is determined as a short-term reference picture, determine that a motion vector of the collocated block is un-available so that the motion vector of the collocated block is not used to predict a motion vector of the current block;   when both of the reference picture of the collocated block and the reference picture of the current block are determined as long-term reference pictures, obtain a temporal motion vector prediction candidate without scaling the motion vector of the collocated block;   generate information about a long-term reference picture for determining whether the reference picture of the current block is the long-term reference picture;   determine a motion vector predictor of the current block from among motion vector prediction candidates, comprising the temporal motion vector prediction candidate; and   generate prediction information of a candidate block as prediction information of the current block that indicates the candidate block used to derive the motion vector predictor of the current block,   wherein the scaling is based on a ratio of a distance (Td) between the collocated picture and the reference picture of the collocated block and a distance (Tb) between the current picture and the reference picture of the current block.   
     
     
         3 . A method for encoding an image, the method comprising:
 generating a collocated index for determining a collocated picture from among pictures restored before a current picture;   when one of a reference picture of a collocated block included in the collocated picture and a reference picture of a current block is determined as a long-term reference picture, and the other one of the reference picture of the collocated block and the reference picture of the current block is determined as a short-term reference picture, determining that a motion vector of the collocated block is un-available so that the motion vector of the collocated block is not used to predict a motion vector of the current block;   when both of the reference picture of the collocated block and the reference picture of the current block are determined as long-term reference pictures, obtaining a temporal motion vector prediction candidate without scaling the motion vector of the collocated block;   generating information about a long-term reference picture for determining whether the reference picture of the current block is the long-term reference picture;   determining a motion vector predictor of the current block from among motion vector prediction candidates, comprising the temporal motion vector prediction candidate; and   generating prediction information of a candidate block as prediction information of the current block that indicates the candidate block used to derive the motion vector predictor of the current block,   wherein the scaling is based on a ratio of a distance (Td) between the collocated picture and the reference picture of the collocated block and a distance (Tb) between the current picture and the reference picture of the current block.   
     
     
         4 . A non-transitory computer-readable storage medium storing a bitstream generated by at least one processor, the bitstream comprising:
 a collocated index for determining a collocated picture from among pictures restored before a current picture;   information about a long-term reference picture for determining whether a reference picture of a current block is a long-term reference picture; and   prediction information of a candidate block as prediction information of the current block that indicates the candidate block used to derive a motion vector predictor of the current block,   wherein when one of a reference picture of a collocated block included in the collocated picture and the reference picture of a current block is determined as the long-term reference picture, and the other one of the reference picture of the collocated block and the reference picture of the current block is determined as a short-term reference picture, a motion vector of the collocated block is determined as un-available so that the motion vector of the collocated block is not used to predict the motion vector of the current block;   when both of the reference picture of the collocated block and the reference picture of the current block are determined as long-term reference pictures, a temporal motion vector prediction candidate is obtained without scaling the motion vector of the collocated block;   wherein the motion vector predictor of the current block is determined from among motion vector prediction candidates, comprising the temporal motion vector prediction candidate   wherein the scaling is based on a ratio of a distance (Td) between the collocated picture and the reference picture of the collocated block and a distance (Tb) between the current picture and the reference picture of the current block.