摘要 |
A prediction difference is generated by performing prediction processing on an input image, quantized data is generated by performing frequency conversion processing and quantization processing on the generated prediction difference, an encoded stream is generated by variable-length encoding the generated quantized data, and in the frequency conversion processing and quantization processing, on the basis of whether or not an intra-screen prediction block is included in a block group comprising a predetermined number of multiple blocks, the size of a block unit in which the frequency conversion processing and the quantization processing are performed is changed, regarding the blocks included in the block group. As a result, the amount of codes is reduced and the subjective image quality is improved more suitably. |