PURPOSE: To prevent light and shade irregularity on an output image from occurring by providing a correction means which corrects the dispersion of image data due to discordance between (odd-number)th and (even-number)th black reference data.
CONSTITUTION: Correction processing is applied to a digital signal sent to a shading arithmetic part 209 based on data for shading correction stored in shading line memory 210 in advance, and the digital signal is sent to a conversion table 211, and the output level of the digital signal is corrected to the optimum output level, and it is sent to a gamma conversion part 212. In such a case, ideal density reproducing data is referred to, and a look-up table changing from a density level set as a sample to the ideal reproducing data is calculated for (odd-number)th and (even-number)th pixels respectively by applying arithmetic processing, and it is stored in a conversion table 211, and the image data after shading correction is converted by the conversion table 211, thereby, ideal image data without generating sensitivity difference for (odd-number)th and (even-number)th output can be obtained.