We derive a principled framework for encoding prior knowledge of information coupling between views or camera poses (translation and orientation) of a single scene. While deep neural networks have become the prominent solution to many tasks in computer vision, some important problems n