A means of reducing the data size of a color image that allows for image processing is presented. By decomposing an image into a spatial information part and a spectral information part, the spatial information may be processed without affecting the spectral content of the image. This significantly reduces the amount of time and computational power required to process a color image. To this point, image processors have attempted such a process by decomposing an image into its luminance and chrominance channels and applying image processing techniques to the luminance channel. Because luminance does not contain all of the spatial information, some spatial information is left unprocessed. The method of channel reduction described herein avoids this problem and has a number of uses in digital still photography, digital video and image compression. |