A method includes specifying an image area which is contained in the subject-present image taken and in which nonflatness is less than or equal to a predetermined value, determining whether or not a range of the specified image area in the subject-present image in which the nonflatness is less than or equal to a predetermined value is greater than or equal to a predetermined range, generating an extraction background image used to extract a subject area in the subject-present image containing the subject, from the image area when the range of the image area in which the nonflatness is less than or equal to a predetermined value is determined to be greater than or equal to the predetermined range, and extracting the subject area from the subject-present image based on information on a difference between each pixel in the extraction background image and a corresponding pixel in the subject-present image.