I do it the old fashioned way: I add up the dimensions of each image width times height, then make a new image that encompasses those dimensions, then copy each image into the new one, overlap as necessary, and flatten.