Images at two or more correction levels are output as corrected images for
a given image, and a user is allowed to select a preferred corrected
image. Based on the selection result of the user, an image to be
corrected which is different from the given image is corrected.