FACTORIZATION METHODS FOR STRUCTURE-FROM-MOTION

In this article we present an overview of factorization methods for re covering structure and motion from image sequences. We distinguish the se methods from general nonlinear algorithms primarily by their biline ar formulation in motion and shape parameters. The bilinear formulatio n makes possible powerful and efficient solution techniques including singular value decomposition. We show how factorization methods apply under various affine camera models and under the perspective camera mo del, and then we review factorization methods for various features inc luding points, lines, directional point features and line segments. An extension to these methods enables them to segment and recover motion and shape for multiple independently moving objects. Finally, we illu strate the generality of the factorization methods with two applicatio ns outside structure from motion.