Monday, July 11

MS13
Linear Algebra in Image Processing

4:00 PM - 6:00 PM
Room: Oak Alley - 3rd Floor

The focus of this minisymposium is on the numerical linear algebra models and algorithms that have been developed for tackling challenging problems in image processing. Many problems in image restoration and reconstruction can be formulated as linear systems or least squares problems. Due to the ill-posedness of these inverse problems, however, sophisticated methods are required to accurately compute approximate solutions in the presence of noise. Furthermore, the algorithms need to exploit the underlying matrix structure for computational efficiency. Topics in this minisymposium include super-resolution imaging, efficient matrix representation and factorization techniques for inverse problems, and projection-based regularization algorithms.

Organizer: Misha E. Kilmer
Tufts University

4:00-4:25 Super-Resolution Image Restoration from Blurred Observations

Andy Chin Ko Yau, University of Hong Kong, China

4:30-4:55 Computing the SVD for Large Matrices in Image Restoration

James G. Nagy, Emory University

5:00-5:25 Analysis and Exploitation of Matrix Structure Arising in Linearized Inverse Scattering

Damon Hyde, Eric Miller, and Dana H. Brooks, Northeastern University

5:30-5:55 An Iterative, Projection-Based Algorithm for General Form Tikhonov Regularization

Malena I. Espanol, Tufts University; Per Christian Hansen, Technical University of Denmark, Denmark; Misha E. Kilmer, Tufts University