METHOD FOR COMPRESSING AND DECOMPRESSING IMAGES BY SUBDIVIDING PIXEL COLOR DISTRIBUTIONS

The present invention provides methods and apparatus for representing visual images in compressed digital form. A series of digital images is provided to a random access memory and compressed one frame at a time. The compressor of the present invention divides a frame into a predefined number of blocks. Each block is tested to determine whether the block has a homogeneous color distribution. If the block has a homogeneous color distribution as defined by a selected homogeneity test, each pixel in that block is encoded with a single color value. If the block does not have a homogeneous color distribution, the block is split into two distributions, one distribution with a greater luminance than the average luminance of the first distribution and one distribution with a smaller luminance than the average luminance of the first distribution. These distributions are separately tested for homogeneity and the same process is repeated until the region has been resolved into homogeneous distributions. Distributions are then merged to check whether the merged distribution is homogeneous. After merging, pixels in homogeneous distributions are encoded as a single color.