Enabling flexible and appropriate response even if a stereoscopic output is inappropriate or not required when generating a stereoscopic viewing image from a three-dimensional image, in which a stereoscopic viewing image generation unit (11) generates, using a three-dimensional medical image representing a subject as input, a stereoscopic viewing image for stereoscopic output of the subject based on a given image generation condition, and a non-stereoscopic viewing image generation unit (12) generates a non-stereoscopic viewing image for non-stereoscopic output equivalent to the stereoscopic output based on the three-dimensional medical image and the image generation condition of the stereoscopic viewing image.