Method and apparatus for coding/decoding image
The present disclosure provides an apparatus for encoding a video including: a prediction unit to generate a predicted block of a current block; a subtraction unit to generate a residual block by subtracting the predicted block from the current block; a transformer to transform the residual block into a frequency transform block by performing different transform methods on different transform units; a quantization unit to quantize the frequency transform block to generate a quantized frequency transform block; and a bitstream generator to encode the quantized frequency transform block into a bitstream.
1. An apparatus for encoding a video, the apparatus comprising:
a bitstream decoder configured to extract information on a size of a transform unit and generate a quantized frequency transform block from a bitstream;
an inverse quantization unit configured to reconstruct a frequency transform block by inversely quantizing the quantized frequency transform block;
an inverse transformer configured to reconstruct a residual block by inversely transforming the frequency transform block, the inverse transformer selecting an inverse transform scheme among a plurality of inverse transform schemes in consideration of the size of the transform unit;
a predictor configured to generate a predicted block in a prediction mode selected among an intra prediction mode and an inter prediction mode; and
an adder configured to reconstruct a current block by adding a first pixel corresponding to the residual block and a second pixel corresponding to the predicted block,
wherein the inverse transformer considers a first transform scheme only when the size of the transform unit corresponds to a transform unit size of 4×4 among a plurality of available transform unit sizes,
wherein the transform unit is divided from a coding unit in a quad tree, and the bitstream decoder decodes a split transform flag, as one of the information on the size of the transform unit, for at least one depth of the quad tree,
wherein the first transform scheme corresponds to a transform scheme in which each of four output values {X′( 0 ), X′( 1 ), X′( 2 ), X′( 3 )} is generated by using at least two of four input values {X( 0 ), X( 1 ), X( 2 ), X( 3 )} multiplied by coefficients corresponding to each of the input values,
wherein the transform scheme in which each of four output values {X′( 0 ), X′( 1 ), X′( 2 ), X′( 3 )} is generated corresponds to a transform in which,
a first group of the coefficients used for generating the first output value X′( 0 ) includes a coefficient for the X( 0 ), a coefficient for the X( 1 ) and a coefficient for the X( 2 ), all having positive signs,
a second group of the coefficients used for generating the second output value X′( 1 ) includes a coefficient for the X( 0 ) and a coefficient for the X( 1 ), both having positive signs,
a third group of the coefficients used for generating the third output value X′( 2 ) includes a coefficient for the X( 1 ) and a coefficient for the X( 2 ), both having negative signs, and
a fourth group of the coefficients used for generating the fourth output value X′( 3 ) included a coefficient for the X( 1 ), having a negative sign, and a coefficient for the X( 2 ), having a positive sign.
2. The apparatus of claim 1 , wherein the inverse transformer selects the inverse transform scheme in consideration of the prediction mode corresponding to the predicted block together with the size of the transform unit.