3D object to 2D image invariance technology for visualization problems

In this paper, we describe applications of a startling recent approach to image recognition and understanding, explore some of the benefits, and list potential visualization applications of this technology. The basis of this work consists of a mathematical formalism which allows one to associate a particular three-dimensional representation of an object or landscape with any of its two-dimensional image views. In contrast to other work in this area, these techniques do not require knowledge of camera parameters (focal length, image scale, etc.) and require only a single image of a given scene. These methods do not rely on trigonometry or off-board information sources (e.g. GPS). They work from image features which are general and common (i.e. lines and points). These properties make the technology particularly desirable for applications such as navigation, cueing, targeting, and image "patching". We will begin with an overview of the geometric theory. We will then discuss what these discoveries allow us to do. Finally we will point to some specific capabilities shortfalls in the military today to which this work could be addressed.