Method and apparatus for encoding/decoding images using adaptive motion vector resolution

Abstract

The present disclosure relates to a method and apparatus for improving the encoding efficiency by adaptively changing the resolution of the motion vector in the inter prediction encoding and inter prediction decoding of a video. The video encoding/decoding apparatus includes: a video encoder for determining a motion vector resolution of each area or motion vector of a video and performing an inter prediction encoding of the video by using a motion vector according to the motion vector resolution determined for each area or motion vector; and a video decoder for reconstructing a resolution by extracting resolution information from a bitstream, and then performing an inter prediction decoding by using a motion vector according to the motion vector resolution of each reconstructed area or motion vector.

1 . A video decoding apparatus for inter-predicting a current block to be decoded, the apparatus comprising:
 a motion vector difference decoder configured to reconstruct a motion vector difference of the current block from a bitstream,   a motion vector resolution decoder configured to decode, from the bitstream, resolution information specifying one among a plurality of motion vector resolutions which respectively indicate resolutions of motion vector differences, and identify the resolution of the motion vector difference of the current block by using the resolution information,   an inter predictor configured to
 derive a motion vector predictor by converting a motion vector candidate derived from a pre-decoded neighboring block of the current block such that the converted motion vector candidate has the same resolution as the resolution of the motion vector difference identified based on the resolution information, 
 reconstruct a current motion vector of the current block by adding the motion vector difference to the motion vector predictor, and 
 predict the current block by interpolating pixels in a reference picture of the current block using the current motion vector. 
   
     
     
         2 . The apparatus of  claim 1 , wherein the plurality of motion vector resolutions includes at least two of a 1 pixel precision, a 1/2 fractional pixel precision, a 1/4 fractional pixel precision, or a 1/8 fractional pixel precision. 
     
     
         3 . The apparatus of  claim 1 , wherein the motion vector resolution decoder is configured to decode the resolution information by
 decoding, from a sequence parameter set of the bitstream, a resolution change flag indicating whether resolutions of motion vector differences of blocks referring to the sequence parameter set are adaptively determined or not, and   decoding resolution identification information of the current block for indicating one among the plurality of motion vector resolutions, when the resolution change flag indicates that the resolutions of the motion vector differences are adaptively determined,   wherein the motion vector resolution decoder is configured to identify the resolution of the motion vector difference of the current block as a motion vector resolution which is indicated by the resolution identification information among the plurality of motion vector resolutions.   
     
     
         4 . The apparatus of  claim 3 , wherein the motion vector resolution decoder is configured to identify the resolution of the motion vector difference of the current block as a pre-defined default motion vector resolution without decoding the resolution identification information of the current block, when the resolution change flag indicates that the resolutions of the motion vector differences are not adaptively determined. 
     
     
         5 . The apparatus of  claim 4 , wherein the default motion vector resolution is a 1/4 fractional pixel precision. 
     
     
         6 . The apparatus of  claim 1 , wherein the inter predictor is configured to
 determine an interpolation filter depending on the resolution of the motion vector difference which is identified based on the resolution information, and   interpolate the pixels in the reference picture using the determined interpolation filter.   
     
     
         7 . The apparatus of  claim 6 , wherein an interpolation filter used in the resolution of the motion vector difference equal to a 1/2 fractional pixel precision is different from those used in the other resolutions. 
     
     
         8 . The apparatus of  claim 1 , wherein the inter predictor is configured to perform a rounding operation for the motion vector candidate and then convert the rounded motion vector candidate.