ABSTRACT
In this presentation, the restoration of images blurred by atmospheric turbulence is examined. The proposal uses a new class of approximations to blurring operators representing Gaussian blur. The Toeplitz matrix representing the blur is transformed into a Cauchy-like (CL) matrix using the FFT. In addition to the CL structure, the transformed matrix has a rank structure. In particular, the off-diagonal blocks have low rank. This class of matrices can be approximated quickly, and the structure can be exploited for fast image restoration.
- Chandrasekaran, S., Gu, M., Sun, X., Xia, J., and Zhu, J. 2007. A superfast algorithm for toeplitz systems of linear equations. SIAM Journal on Matrix Analysis and Applications 29, 4, 1247--1266. Google ScholarDigital Library
- Hansen, P. C., and Jensen, T. K. 2008. Noise propagation in regularizing iterations for image deblurring. Electronic Transactions on Numerical Analysis 31, 204--220.Google Scholar
- Nagy, J. G., Plemmons, R. J., and Torgersen, T. C. 1996. Iterative image restoration using approximate inverse preconditioning. IEEE Transactions on Image Processing 5, 7 (July), 1151--1162. Google ScholarDigital Library
Index Terms
- Deblurring with rank-structured inverse approximations
Recommendations
Compressing Rank-Structured Matrices via Randomized Sampling
Randomized sampling has recently been proven a highly efficient technique for computing approximate factorizations of matrices that have low numerical rank. This paper describes an extension of such techniques to a wider class of matrices that are not ...
Structured Hölder Condition Numbers for Multiple Eigenvalues
The sensitivity of a multiple eigenvalue of a matrix under perturbations can be measured by its Hölder condition number. Various extensions of this concept are considered. A meaningful notion of structured Hölder condition numbers is introduced, and it ...
Stability Issues in the Factorization of Structured Matrices
This paper provides an error analysis of the generalized Schur algorithm of Kailath and Chun [SIAM J. Matrix Anal. Appl., 15 (1994), pp. 114--128]---a class of algorithms which can be used to factorize Toeplitz-like matrices, including block-Toeplitz ...
Comments