Efficient Coding of Computer Generated Images Utilizing FIFO Buffers

The technique of reversible compression coding for images of computer graphics (CG) is important in graphic systems where it is inconvenient to predetermine distortion allowances, such as image data base systems. The authors have already proposed Structure Runlength Coding (SRC) and Boundary Detection Coding (BDC), which are reversible coding schemes for CG images. Those schemes use the characteristics of CG images of having many uniformly colored areas. They compress the configuration of uniformly colored areas rather efficiently. Color signals are, however, coded without any compression measures : if a given image has twenty-four bits for a pixel, the codes use the same bit pattern for the colors of areas. The present paper proposes an FIFO buffer scheme which compresses color signals efficiently. Using an FIFO buffer, some colors in an image can be coded indirectly with the addresses of the colors in the buffer. The FIFO scheme will be built in SRC scheme and BDC scheme, making both of the coding schemes more efficient. Using a buffer of suitable sizes, the sample images in our experiment are compressed about 10% in average to the original size, while the com presion ratio of the prototype schemes are around 15%.