| 摘要 |
The image encoding method according to an embodiment comprises a step (112) of establishing, on the basis of relationships defined beforehand in accordance with the predicted image generating manners of intra-prediction modes, a combination of vertical and horizontal transformation matrices that are to be encoded and correspond to the intra-prediction modes. This combination is established by combining any ones of a plurality of transformation matrices including first transformation matrices and further including second transformation matrices the coefficient collection densities of which are higher than those of the first transformation matrices in a case where prediction errors of the intra-prediction modes in each of which a reference pixel group on at least one line is referenced to generate an intra-predicted image are subjected to a one-dimensional orthogonal transformation in a direction orthogonal to the line of the reference pixel group. The image encoding method further comprises a step (102) of performing, by use of the established vertical and horizontal transformation matrices, vertical and horizontal transformations for the prediction errors, thereby obtaining transformation coefficients. |
