An image processing device is provided including: an acquiring unit that acquires position data expressing a position of plural reference images provided on a reference medium; a scanning unit that scans the plural reference images; and a determination unit. The determination unit derives the sum of positional misalignment amounts of corresponding reference images, between the position of reference images expressed by image data obtained by the scanning unit, and the position of reference images expressed by position data acquired by the acquiring unit. The determination unit executes alignment of the reference images using a common correction coefficient predetermined so as to give the smallest sum of the positional misalignment amounts. Then the determination unit determines a dilation amount for image dilation on image data such that the dilation amount increases as the sum of the positional misalignment amounts increases.