An image processing apparatus includes a size reducing unit that reduces sizes of a plurality of continuous images, which are still images obtained by continuously capturing images of a moving object, to thereby generate reduced continuous images; a mask generating unit that extracts moving object regions from the reduced continuous images, to thereby generate reduced moving object extraction mask images; a size restoring unit that enlarges the reduced moving object extraction mask images to the same size as original sizes of the continuous images that are not reduced by the size reducing unit, to thereby generate moving object extraction mask images; and a combining unit that extracts the moving object regions from the continuous images by using the moving object extraction mask images to thereby obtain moving object images, and combines the moving object images in a predetermined one of the continuous images.