Method for generating prediction block in amvp mode

Abstract

Discussed is a video decoding apparatus including a merge mode motion information decoding unit that decodes motion information if a motion information encoding mode of a current block is a merge mode, a prediction block generation unit that generates a prediction block of the current block using the decoded motion information, and a residual block decoding unit that generates a quantization block by inversely scanning a residual signal to dequantize the quantization block using a quantization step size, and to generate a residual block by inversely transforming the dequantization block.

1 . A video decoding apparatus, comprising:
 a merge mode motion information decoding unit configured to decode motion information if a motion information encoding mode of a current block is a merge mode;   a prediction block generation unit configured to generate a prediction block of the current block using the decoded motion information; and   a residual block decoding unit configured to generate a quantization block by inversely scanning a residual signal, to dequantize the quantization block using a quantization step size, and to generate a residual block by inversely transforming the dequantization block,   wherein the quantization step size is generated using an average value of effective two of a left quantization step size, an upper quantization step size, and a previous quantization step size of a current coding unit, and   wherein the quantization step size is decoded using an effective quantization step size if only one of the left quantization step size, the upper quantization step size, and the previous quantization step size is effective.   
     
     
         2 . The video decoding apparatus according to  claim 1 , wherein:
 the merge mode motion information decoding unit comprises a spatial merge candidate derivation unit and a temporal merge candidate configuration unit, and   the temporal merge candidate configuration unit configures a temporal merge candidate using a reference picture index of the temporal merge candidate and a motion vector of the temporal merge candidate.   
     
     
         3 . The video decoding apparatus according to  claim 2 , wherein the reference picture index of the temporal merge candidate is set as a reference picture index of one of blocks contiguous to the current block or set to 0. 
     
     
         4 . The video decoding apparatus according to  claim 3 , wherein:
 the temporal merge candidate configuration unit determines a temporal merge candidate picture and determines a temporal merge candidate block within the temporal merge candidate picture, and   the temporal merge candidate picture is differently determined depending on a slice type.   
     
     
         5 . The video decoding apparatus according to  claim 4 , wherein a first candidate block or a second candidate block is determined as the temporal merge candidate block depending on a position of the current block. 
     
     
         6 . The video decoding apparatus according to  claim 5 , wherein:
 the first candidate block is a left-below corner block of a block corresponding to a current prediction unit within the temporal merge candidate picture, and   the second candidate block is a block comprising a right-above pixel at a center of the block corresponding to the current prediction unit.   
     
     
         7 . The video decoding apparatus according to  claim 1 , further comprising:
 a merge candidate generation unit configured to generate a merge candidate if a number of effective merge candidates of the current block is smaller than a predetermined number,   wherein the merge candidate generation unit generates the merge candidate having a motion vector of 0 and a reference picture index of 0.   
     
     
         8 . The video decoding apparatus according to  claim 7 , wherein the merge candidate generation unit generates the merge candidate having a number smaller than or equal to the predetermined number if the merge candidate is generated by combining motion information of predetermined effective merge candidates.