Based on a plurality of detection data sets acquired by a data acquisition unit and correspond to respective tube voltages, an image generator generates a plurality of reference substance image data sets targeting respective reference substances contained in a subject. The image generator generates a correction data set for each of the plurality of tube voltages by applying, to a detection data set corresponding to each of the tube voltages, a correction coefficient for suppressing a discrepancy of an actual energy spectrum of the X-rays detected by an X-ray detector from a predetermined X-ray energy spectrum, and generates the plurality of reference substance image data sets based on a plurality of correction data sets corresponding to the respective tube voltages.
