| 摘要 |
The image processing technique of the present invention prevents the deterioration of picture quality due to quantization errors caused by conversion of the tone number, which allocates image data to lattice points previously set in a color space. Pre-conversion of the tone number allocates original color image data to peripheral lattice points among the lattice points set by dividing the color space into lattices. The allocation of the original color image data to the peripheral lattice points causes quantization errors. The allocation, however, causes the mean error to be not greater than a predetermined value, in order to output desirable colors in a resulting image. The color space may be divided into smaller areas in a low-density range. This minimizes the deterioration of picture quality due to the quantization errors. A smoothing process carried out after the pre-conversion of the tone number relieves the effect of the quantization errors and further improves the picture quality.
|
