Abstract | ||
---|---|---|
Regularization by denoising (RED), as recently proposed by Romano, Elad, and Milanfar, is powerful image-recovery framework that aims to minimize an explicit regularization objective constructed from a plug-in image-denoising function. Experimental evidence suggests that the RED algorithms are a state of the art. We claim, however, that explicit regularization does not explain the RED algorithms. In particular, we show that many of the expressions in the paper by Romano
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</italic>
hold only when the denoiser has a symmetric Jacobian, and we demonstrate that such symmetry does not occur with practical denoisers such as nonlocal means, BM3D, TNRD, and DnCNN. To explain the RED algorithms, we propose a new framework called Score-Matching by Denoising (SMD), which aims to match a “score” (i.e., the gradient of a log-prior). We then show tight connections between SMD, kernel density estimation, and constrained minimum mean-squared error denoising. Furthermore, we interpret the RED algorithms from Romano
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</italic>
and propose new algorithms with acceleration and convergence guarantees. Finally, we show that the RED algorithms seek a consensus equilibrium solution, which facilitates a comparison to plug-and-play ADMM. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/TCI.2018.2880326 | IEEE Transactions on Computational Imaging |
Keywords | DocType | Volume |
Noise reduction,Jacobian matrices,Estimation,Imaging,Convergence,Optimization,Kernel | Journal | abs/1806.02296 |
Issue | ISSN | Citations |
1 | 2333-9403 | 5 |
PageRank | References | Authors |
0.40 | 25 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Edward T. Reehorst | 1 | 5 | 0.40 |
Philip Schniter | 2 | 1620 | 93.74 |