| 摘要 |
A method of color image compression processing provides compensation for lossy luminance compression (30) by adjusting (34) the red and blue color components using a lossy luminance compression factor equal to the difference between compressed and decompressed luminance. The compensated color components are then used to compute (42) chrominance color differences. Chrominance is range compressed (44) by bit-reduction using a non-linear range compression function that minimizes compression for small chrominance magnitudes, where the eye is more sensitive to a loss of color fidelity, while maximizing compression in the color saturation regions where the eye is less sensitive. After luminance and chrominance decompression (60) and reconversion to the RBG color components (70), compensation for red and blue color shift (such as due to round-off errors) is accomplished by multiplying (80) the green component (which is calculated from the color-shifted red and blue components) by a color shift compensation factor that shifts green in the opposite direction from the shift in the red and blue components.
|
