An image processing apparatus comprising an image producing unit 101 for producing an axial image of a body part to be imaged including an aorta and an esophagus; a map generating unit 102 for generating a map M2 for locating a region in which a probability that the aorta lies is high in the axial image; a detecting unit 103 for detecting a temporary position of the aorta based on the map M2; and a deciding unit 104 for making a decision on whether or not the temporary position of the aorta falls within the region of the aorta in the axial image based on a distribution model DM containing information representing a reference position (xe, ye) of the esophagus and information representing a range over which the aorta distributes relative to the reference position (xe, ye) of the esophagus, and on the map M2.