When image data are input from a memory card or the like, a printing apparatus detects face region captured in these image data. Where the size and position of a face region fall within permitted ranges therefor set in advance in a permitted range table derived via sense evaluation testing, the input image data is selected as a printing target candidate. If the face region size and position do not each fall within their permitted range, however, the image data is not selected as a printing target. As a result, image data suitable for printing can be efficiently selected as printing targets from image data that contain human faces.