The present invention provides a technique for enabling even a person having no expert knowledge to adjust the direction of an optical axis easily with high precision. In a three-dimensional image processing apparatus having plural image capturing apparatuses, optical center of an image from an image capturing apparatus is stored for each of the image capturing apparatuses. A mark indicative of the optical center is superimposed on an image obtained from the image capturing apparatus. By referring to the mark for each of the image capturing apparatuses, the orientation of each of the image capturing apparatuses is adjusted so that all of the marks overlay on the same point on an object.