Adaptive transformation of residual blocks depending on the intra prediction mode

摘要

In an embodiment a moving-picture decoding apparatus comprises:
an entropy decoder configured to obtain transformation coefficients from coded data obtained by coding processes according to prediction directions;
an order controller configured to rearrange the transformation coefficients according to a scanning order corresponding to a prediction direction of a target block from among scanning orders predetermined for the prediction directions;
a classifying unit configured to classify a mode applied to the target block into a predetermined first mode or a predetermined second mode;
a first inverse orthogonal transformer configured to obtain a predictive residual signal by subjecting the rearranged transformation coefficients to an inverse orthogonal transformation by use of a first transformation basis common to the prediction directions if the applied mode is classified into the first mode;
a second inverse orthogonal transformer configured to obtain a predictive residual signal by subjecting the rearranged transformation coefficients to the inverse orthogonal transformation by use of a second transformation basis different from the first transformation basis if the applied mode is classified into the second mode; and
a generator configured to generate a decoded image signal by use of the predictive residual signal,
wherein the inverse orthogonal transformation differs between a luminance signal and a color-difference signal, the inverse orthogonal transformation being performed by the first inverse orthogonal transformer and the second inverse orthogonal transformer.