There would be no difference if you decide on the processing algorithm first and then generate images from an input. But in my mind image processing would also include post production of an image. Computer vision is a set of techniques that could be used in an algorithmic camera.