A method of compressing a color image is provided. The color image comprises
data for a plurality of pixels. The method includes the step of obtaining red,
green and blue pixel values of an object of interest in the image. A calculation
is made of the complement of the red, green and blue values of the object of interest.
Transformation coefficients are calculated which transform the complements of red,
green and blue values of the object of interest into representations in a transformation
color space. The transformation coefficients are applied to all the pixels in the
image to thereby obtain a transformed data set representing the image having components
along three mutually orthogonal axes (A, B and C herein) in a three-dimensional
transformed color space. The transformed data set is scaled in accordance with
the color quantization used in the system; e.g., the A, B and C values are between
0 and 255 for an 8 bit quantization. A compression algorithm, e.g., a loss less
algorithm such as WINZip or LZW is applied to at least two components of the transformed
data set to thereby produce output data representing a compression of the image.