* Extension should be able to use computer vision for every picture on page if user asked. * If there is "cloud" in tags of recognized objects, paste the butt image on the picture in that place.