摘要 |
<p>PROBLEM TO BE SOLVED: To obtain a high compression rate while keeping excellent image quality by selecting one of a plurality of compression systems. SOLUTION: A block dividing section 11 divides each version of an original image (a CMYK image in this case) being a compression object into macro blocks. A DCT transform section 12 applies DCT transform being a kind of orthogonal transforms to each block to breakdown each block into frequency components. The frequency component is fed to a quantization section 13, where the frequency component is linearly quantized with a different step size from each position of coefficient by a quantization table in the unit of blocks. The quantized coefficient is entropy-coded 14 and compressed with a variable length. A decoding section 15 at a decoder decodes the compressed code of each version into a quantized value through a reverse path to compression and an inverse quantization section 16 restores the quantized value into the frequency coefficient by a table of each version. Then an inverse DCT transform section 17 restores the frequency coefficient into density data to generate each version of the CMYK images.</p> |