An image processing system receives an input gradient number m, input image data I(x,y) whose input gradient number is m, and an output gradient number n and calculates output image data O(x,y) having an output gradient number of n on the basis of the following formula (1). O(x,y)=int[α+d(xd,yd) / (gmax+1)] (1) When α is an integer, the output image data O(x,y) is calculated on the basis of a formula obtained by substituting the calculated α′ for α in the above formula (1), whereas when α is not an integer, the output image data O(x,y) is calculated on the basis of the above formula (1).