In a recording mode, when a shutter release button is pressed with a panning mode being selected as a result of press of a panning-mode button, plural pre-combined images are captured through continuous photographing using an image sensor. After continuous photographing, partial images (moving-subject images) each showing a moving subject which is located differently among the pre-combined images are detected in an image combiner. Then, the plural pre-combined images are combined such that respective positions of the detected moving-subject images in the pre-combined images are substantially identical to one another, to create one frame of composite image. In the created composite image, while the moving subject is frozen, a background (objects other than the moving subject) appears to flow because of differences in positional relationship between the moving subject and the background among the pre-combined images.