Share Email Print
cover

Proceedings Paper

The multiscale structure of non-differentiable image manifolds
Author(s): Michael B. Wakin; David L. Donoho; Hyeokho Choi; Richard G. Baraniuk
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

In this paper, we study families of images generated by varying a parameter that controls the appearance of the object/scene in each image. Each image is viewed as a point in high-dimensional space; the family of images forms a low-dimensional submanifold that we call an image appearance manifold (IAM). We conduct a detailed study of some representative IAMs generated by translations/rotations of simple objects in the plane and by rotations of objects in 3-D space. Our central, somewhat surprising, finding is that IAMs generated by images with sharp edges are nowhere differentiable. Moreover, IAMs have an inherent multiscale structure in that approximate tangent planes fitted to ε-neighborhoods continually twist off into new dimensions as the scale parameter ε varies. We explore and explain this phenomenon. An additional, more exotic kind of local non-differentiability happens at some exceptional parameter points where occlusions cause image edges to disappear. These non-differentiabilities help to understand some key phenomena in image processing. They imply that Newton's method will not work in general for image registration, but that a multiscale Newton's method will work. Such a multiscale Newton's method is similar to existing coarse-to-fine differential estimation algorithms for image registration; the manifold perspective offers a well-founded theoretical motivation for the multiscale approach and allows quantitative study of convergence and approximation. The manifold viewpoint is also generalizable to other image understanding problems.

Paper Details

Date Published: 17 September 2005
PDF: 17 pages
Proc. SPIE 5914, Wavelets XI, 59141B (17 September 2005); doi: 10.1117/12.617822
Show Author Affiliations
Michael B. Wakin, Rice Univ. (United States)
David L. Donoho, Stanford Univ. (United States)
Hyeokho Choi, Rice Univ. (United States)
Richard G. Baraniuk, Rice Univ. (United States)


Published in SPIE Proceedings Vol. 5914:
Wavelets XI
Manos Papadakis; Andrew F. Laine; Michael A. Unser, Editor(s)

© SPIE. Terms of Use
Back to Top