To provide an image processing technology by which the image of an object will not be blurred or duplicated even if the object which has moved into a stitched portion of the image, at the time of the composition of the object images acquired from dynamic images of photographed images or the like.

When two or more object images (frame images) are joined together, a relative positional relationship between the two or more object images is decided, two or more frame images to be overlapped are identified, the overlapped portion of the two or more identified frame images is divided into one or more blocks which compart the same area in each of the overlapped frame images. For each block, the movement of an photographic subject existing in the block is determined. Concerning a block including a photographic subject which has been decided as the one having moved over a given distance or more, one frame image, among two or more frame images to be overlapped, is used, and concerning other blocks, two or more frame images to be overlapped are used, thus generating a still image. Consequently, blurring at the overlapped portion is prevented.