A storage (240) of an image processing device (100) stores a background image (2410) and a first image (2420) and a second image (2430) that are foreground images with respect to the background image (2410), and a first basic mask image (the first basic mask image (2450) formed from pixels having a pixel value of ‘1’ indicating opacity and pixels having a pixel value of ‘0’ indicating transparence. An acquirer (250) acquires each of these images. An image generator (220) generates a first intermediate image by compositing the first image (2420) onto the background image (2410) only in pixels that are in the same positions as the pixels in the first basic mask image (2450) having the pixel value of ‘1’. The image generator (220) then generates a first composite image by overwriting the second image (2430) onto the first intermediate image only in the pixels that are in the same positions as the pixels in the first basic mask image (2450) having the pixel value of ‘1’.