2016
DOI: 10.1137/15m1042280
|View full text |Cite
|
Sign up to set email alerts
|

Off-the-Grid Recovery of Piecewise Constant Images from Few Fourier Samples

Abstract: We introduce a method to recover a continuous domain representation of a piecewise constant two-dimensional image from few low-pass Fourier samples. Assuming the edge set of the image is localized to the zero set of a trigonometric polynomial, we show the Fourier coefficients of the partial derivatives of the image satisfy a linear annihilation relation. We present necessary and sufficient conditions for unique recovery of the image from finite low-pass Fourier samples using the annihilation relation. We also … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

1
154
0
1

Year Published

2016
2016
2023
2023

Publication Types

Select...
8

Relationship

2
6

Authors

Journals

citations
Cited by 114 publications
(156 citation statements)
references
References 69 publications
1
154
0
1
Order By: Relevance
“…For better readability, the theory here is outlined by assuming 1-D signals, but the principle can be extended for multidimensional signals [22].…”
Section: B Transform Domain Sparsity and Low-rankness In Weighted K-mentioning
confidence: 99%
See 2 more Smart Citations
“…For better readability, the theory here is outlined by assuming 1-D signals, but the principle can be extended for multidimensional signals [22].…”
Section: B Transform Domain Sparsity and Low-rankness In Weighted K-mentioning
confidence: 99%
“…H (l ĝ i ) ∈ C (n −d+1)×d are used in (53) instead of H c (l ĝ i )), then due to the special structure of the Hankel matrix, Y v becomes the transpose of Y h ; accordingly, by interchanging the role of n − d + 1 and d, one can make the rank of Y v equal to that of Y h . However, the theoretical analysis of the concatenated Hankel matrices without wrap around property turns out to be quite involved due to the boundary condition and data truncation; and even more, it is even not necessary in accelerated MR except for the super-resolution imaging (see [22]), because the image should be recovered on a fixed grid rather than on a continuum which makes the cardinal spline model more appropriate. Under this condition, the performance of the resulting rank minimization for the vertical stacking approach becomes deteriorated as will be shown in Discussion (see Supplement Material).…”
Section: ) Low-rank Matrix Completionmentioning
confidence: 99%
See 1 more Smart Citation
“…The linear dependencies between the Fourier coefficients exploited the in structured low-rank matrix priors result from a variety of assumptions, including continuous domain analogs of sparsity [3], [6], [8], [9], correlations in the locations of the sparse coefficients in space [8], [9], multi-channel sampling [2], [10], [11], or smoothly varying complex phase [4]. For example, the LORAKS framework [4] capitalized on the sparsity and smooth phase of the continuous domain image using structured low-rank priors, which offers improved reconstructions over conventional ℓ 1 recovery.…”
Section: Introductionmentioning
confidence: 99%
“…Vetterli et al (2002) proposed, for example, a sampling scheme permitting the exact recovery of a stream of Dirac from a few Fourier series coefficients. The framework has since been applied successfully in other fields, and extended to 2D signals as well as noisy measurements (Maravić & Vetterli 2005;Shukla & Dragotti 2007;Pan et al 2014;Ongie & Jacob 2016). Having been originally designed to work only with equally spaced Fourier samples as input, Pan et al (2017b) extended the FRI framework to cases with non-uniform samples (as is the case in radio interferometry).…”
Section: Introductionmentioning
confidence: 99%