Do I have to copy all pixels over to make a sub image?

Processing or even SFML made it easy to grab a sub image out of an old image without having to copy over each pixel, I cant seem to find any documentation on easy sub images, why is this?