Input image data (255, 100, 50) is first converted into output image data (240, 110, 40) based on the color conversion properties of the input end device and the output device, and is subsequently converted into final output image data (255, 115, 40). The latter conversion satisfies the relationship of (240−40):(110−40)=(255−40):(115−40). By changing the highest-gradation-level color in the output image data into the maximum gradation value (255), the complementary color component is made zero in the output results of the output device, thereby reducing granularity Since the middle-gradation-level color is corrected, variations in color hue are reduced.