Image compression apparatus and image compression program

An image compression apparatus for compressing image data that has a luminance signal and chrominance signals, comprising:

• a distortion amount setting part that sets a target distortion amount of the image data;
• a subsampling rate setting part that sets a subsampling rate of the chrominance signals in accordance with the target distortion amount;
• a subsampling processing part that performs subsampling processing only for the chrominance signals in the image data at the subsampling rate;
• and
• a compression encoding part that subjects the subsampling-processed image data to compression encoding in accordance with the target distortion amount, the compression encoding part including:
• an image transformation part which subjects the image data to sub-band decomposition into frequency regions, and produces transformation coefficients,
• a quantizing part that quantizes the transformation coefficients, and
• an encoding part that encodes the quantized transformation coefficients,
• wherein the subsampling setting part sets the subsampling ate of the chrominance signals at a coarse value when the target distortion amount is large and sets the subsampling rate of the chrominance signals at a fine value when the target distortion amount is small.

A method for compressing image data that has a luminance signal and chrominance signals, comprising:

• setting a target distortion amount of the image data;
• setting a subsampling rate of the chrominance signals in accordance with the target distortion amount;
• performing subsampling processing only for the chrominance signals in the image data at the subsampling rate;
• subjecting the subsampling-processed image data to compression encoding in accordance with the target distortion amount, the compression encoding including:
• subjecting the image data to sub-band decomposition into frequency regions, and producing transformation coefficients,
• quantizing the transformation coefficients, and
• encoding the quantized transformation coefficients;
• and
• setting the subsampling rate of the chrominance signals at a coarse value when the target distortion amount is large, and setting the subsampling rate of the chrominance signals at a fine value when the target distortion amount is small.