Video encoding/decoding method and apparatus for motion compensation prediction

Abstract

A video encoding method and apparatus to select one combination, for each block of an input video signal, from a plurality of combinations. Each combination includes a predictive parameter and at least one reference picture number determined in advance for the reference picture. A prediction picture signal is generated in accordance with the reference picture number and predictive parameter of the selected combination. A predictive error signal is generated representing an error between the input video signal and the prediction picture signal. Encoding the predictive error signal, information of the motion vector, and index information indicating the selected combination is included.

1 . A video decoding method for decoding encoded data obtained by subjecting a video image having luminance and two color differences to prediction encoding to the video image, comprising:
 receiving an input of encoded data obtained by encoding, for one or more to-be-decoded blocks, a plurality of combinations comprising (1) a weighting factor for each luminance and for each of two color differences, (2) an offset for each luminance and for each of two color differences, (3) a flag indicating presence or non-presence of the weighting factor and the offset concerning luminance and (4) a flag indicating presence or non-presence of the weighting factors and the offsets concerning the two color differences, and encoding, for a to-be-decoded block, (A) a quantized orthogonal transform coefficient of a prediction error signal concerning luminance and two color differences and (B) an index indicating (a) one combination of the plurality of combinations and (b) a reference image;   deriving the weighting factors and the offsets from the plurality of combinations and the index;   generating a prediction image by multiplying the reference image by the weighting factors and adding the offsets;   generating a prediction error signal by subjecting the quantized orthogonal transform coefficient to inverse quantization and inverse orthogonal transform; and   generating a decoded image by calculating a sum of the prediction error signal and the prediction image;   wherein two possible values of the index correspond to different reference images, and indicate an identical combination of the weighting factors and the offsets, and a value obtained by arithmetic-shifting right one of the two possible values by one bit is identical to a value obtained by arithmetic-shifting right the other value by one bit.   
     
     
         2 . A video decoding apparatus for decoding encoded data obtained by subjecting a video image having luminance and two color differences to prediction encoding to the video image, comprising:
 a receiver configured to receive an input of encoded data obtained by encoding, for one or more to-be-decoded blocks, a plurality of combinations comprising (1) a weighting factor for each luminance and for each of two color differences, (2) an offset for each luminance and for each of two color differences, (3) a flag indicating presence or non-presence of the weighting factor and the offset concerning luminance and (4) a flag indicating presence or non-presence of the weighting factors and the offsets concerning the two color differences, and encoding, for a to-be-decoded block, (A) a quantized orthogonal transform coefficient of a prediction error signal concerning luminance and two color differences and (B) an index indicating (a) one combination of the plurality of combinations and (b) a reference image;   a deriving module configured to derive the weighting factor and the offset from the plurality of combinations and the index;   a first generator configured to generate a prediction image by multiplying the reference image by the weighting factors and adding the offsets;   a second generator configured to generate a prediction error signal by subjecting the quantized orthogonal transform coefficient to inverse quantization and inverse orthogonal transform; and   a third generator configured to generate a decoded image by calculating a sum of the prediction error signal and the prediction image;   wherein two possible values of the index correspond to different reference images, and indicate an identical combination of the weighting factors and the offsets, and a value obtained by arithmetic-shifting right one of the two possible values by one bit is identical to a value obtained by arithmetic-shifting right the other value by one bit.