SherbyNet - /page/wikibook/entry.mu

<!DOCTYPE html>

<title>Compressed sensing</title>

</head>

Compressed sensing

</h1>

</header>

Compressed sensing (also known as compressive sensing, compressive sampling, or sparse sampling) is a <a href="Signal_processing" title="Signal processing">signal processing</a> technique for efficiently acquiring and reconstructing a <a href="Signal_(electronics)" class="mw-redirect" title="Signal (electronics)">signal</a> by finding solutions to <a href="Underdetermined_system" title="Underdetermined system">underdetermined linear systems</a>. This is based on the principle that, through optimization, the sparsity of a signal can be exploited to recover it from far fewer samples than required by the <a href="Nyquist%E2%80%93Shannon_sampling_theorem" title="Nyquist–Shannon sampling theorem">Nyquist–Shannon sampling theorem</a>. There are two conditions under which recovery is possible. The first one is <a href="Sparsity" class="mw-redirect" title="Sparsity">sparsity</a>, which requires the signal to be sparse in some domain. The second one is incoherence, which is applied through the isometric property, which is sufficient for sparse signals.<a href="#cite_note-1">[1]</a><a href="#cite_note-2">[2]</a> Compressed sensing has applications in, for example, <a href="Magnetic_resonance_imaging" title="Magnetic resonance imaging">magnetic resonance imaging</a> (MRI) where the incoherence condition is typically satisfied.<a href="#cite_note-3">[3]</a>

<div class="mw-heading mw-heading2"><h2 id="Overview">Overview</h2></div>

A common goal of the engineering field of <a href="Signal_processing" title="Signal processing">signal processing</a> is to reconstruct a signal from a series of sampling measurements. In general, this task is impossible because there is no way to reconstruct a signal during the times that the signal is not measured. Nevertheless, with prior knowledge or assumptions about the signal, it turns out to be possible to perfectly reconstruct a signal from a series of measurements (acquiring this series of measurements is called <a href="Sampling_(signal_processing)" title="Sampling (signal processing)">sampling</a>). Over time, engineers have improved their understanding of which assumptions are practical and how they can be generalized.

An early breakthrough in signal processing was the <a href="Nyquist%E2%80%93Shannon_sampling_theorem" title="Nyquist–Shannon sampling theorem">Nyquist–Shannon sampling theorem</a>. It states that if a <a href="Real_number" title="Real number">real</a> signal's highest frequency is less than half of the sampling rate, then the signal can be reconstructed perfectly by means of <a href="Whittaker%E2%80%93Shannon_interpolation_formula" title="Whittaker–Shannon interpolation formula">sinc interpolation</a>. The main idea is that with prior knowledge about constraints on the signal's frequencies, fewer samples are needed to reconstruct the signal.

Around 2004, <a href="Emmanuel_Cand%C3%A8s" title="Emmanuel Candès">Emmanuel Candès</a>, <a href="Justin_Romberg" class="mw-redirect" title="Justin Romberg">Justin Romberg</a>, <a href="Terence_Tao" title="Terence Tao">Terence Tao</a>, and <a href="David_Donoho" title="David Donoho">David Donoho</a> proved that given knowledge about a signal's <a href="Sparsity" class="mw-redirect" title="Sparsity">sparsity</a>, the signal may be reconstructed with even fewer samples than the sampling theorem requires.<a href="#cite_note-4">[4]</a><a href="#cite_note-Donoho-5">[5]</a> This idea is the basis of compressed sensing.

<div class="mw-heading mw-heading2"><h2 id="History">History</h2></div>

Compressed sensing relies on <a href="Lp_space" title="Lp space"><math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy"></a> techniques, which several other scientific fields have used historically.<a href="#cite_note-6">[6]</a> In statistics, the <a href="Least_squares" title="Least squares">least squares</a> method was complemented by the <a href="Lp_norm" class="mw-redirect" title="Lp norm"><math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy">-norm</a>, which was introduced by <a href="Pierre-Simon_Laplace" title="Pierre-Simon Laplace">Laplace</a>. Following the introduction of <a href="Linear_programming" title="Linear programming">linear programming</a> and <a href="George_Dantzig" title="George Dantzig">Dantzig</a>'s <a href="Simplex_algorithm" title="Simplex algorithm">simplex algorithm</a>, the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy">-norm was used in <a href="Computational_statistics" title="Computational statistics">computational statistics</a>. In statistical theory, the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy">-norm was used by <a href="George_W._Brown_(academic)" class="mw-redirect" title="George W. Brown (academic)">George W. Brown</a> and later writers on <a href="Median-unbiased_estimator" class="mw-redirect" title="Median-unbiased estimator">median-unbiased estimators</a>. It was used by <a href="Peter_J._Huber" title="Peter J. Huber">Peter J. Huber</a> and others working on <a href="Robust_statistics" title="Robust statistics">robust statistics</a>. The <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy">-norm was also used in signal processing, for example, in the 1970s, when seismologists constructed images of reflective layers within the earth based on data that did not seem to satisfy the <a href="Nyquist%E2%80%93Shannon_sampling_theorem" title="Nyquist–Shannon sampling theorem">Nyquist–Shannon criterion</a>.<a href="#cite_note-7">[7]</a> It was used in <a href="Matching_pursuit" title="Matching pursuit">matching pursuit</a> in 1993, the <a href="Lasso_regression" class="mw-redirect" title="Lasso regression">LASSO estimator</a> by <a href="Robert_Tibshirani" title="Robert Tibshirani">Robert Tibshirani</a> in 1996<a href="#cite_note-8">[8]</a> and <a href="Basis_pursuit" title="Basis pursuit">basis pursuit</a> in 1998.<a href="#cite_note-9">[9]</a>

At first glance, compressed sensing might seem to violate <a href="Nyquist%E2%80%93Shannon_sampling_theorem" title="Nyquist–Shannon sampling theorem">the sampling theorem</a>, because compressed sensing depends on the <a href="Sparse_matrix" title="Sparse matrix">sparsity</a> of the signal in question and not its highest frequency. This is a misconception, because the sampling theorem guarantees perfect reconstruction given sufficient, not necessary, conditions. A sampling method fundamentally different from classical fixed-rate sampling cannot "violate" the sampling theorem. Sparse signals with high frequency components can be highly under-sampled using compressed sensing compared to classical fixed-rate sampling.<a href="#cite_note-10">[10]</a>

<div class="mw-heading mw-heading2"><h2 id="Method">Method</h2></div>

<div class="mw-heading mw-heading3"><h3 id="Underdetermined_linear_system">Underdetermined linear system</h3></div>

An <a href="Underdetermined_system" title="Underdetermined system">underdetermined system</a> of linear equations has more unknowns than equations and generally has an infinite number of solutions. The figure below shows such an equation system <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathbf {y} =D\mathbf {x} }">

</mrow>

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathbf {y} =D\mathbf {x} }</annotation>

</semantics>

</math><img src="./536cb1951fc4c2034d40aa05e02c97ff537a06a8.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:7.845ex; height:2.509ex;" alt="{\displaystyle \mathbf {y} =D\mathbf {x} }" loading="lazy"> where we want to find a solution for <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathbf {x} }">

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathbf {x} }</annotation>

</semantics>

</math><img src="./32adf004df5eb0a8c7fd8c0b6b7405183c5a5ef2.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.411ex; height:1.676ex;" alt="{\displaystyle \mathbf {x} }" loading="lazy">.

In order to choose a solution to such a system, one must impose extra constraints or conditions (such as smoothness) as appropriate. In compressed sensing, one adds the constraint of sparsity, allowing only solutions which have a small number of nonzero coefficients. Not all underdetermined systems of linear equations have a sparse solution. However, if there is a unique sparse solution to the underdetermined system, then the compressed sensing framework allows the recovery of that solution.

<div class="mw-heading mw-heading3"><h3 id="Solution_/_reconstruction_method">Solution / reconstruction method</h3></div>

Compressed sensing takes advantage of the redundancy in many interesting signals—they are not pure noise. In particular, many signals are <a href="Sparse_matrix" title="Sparse matrix">sparse</a>, that is, they contain many coefficients close to or equal to zero, when represented in some domain.<a href="#cite_note-11">[11]</a> This is the same insight used in many forms of <a href="Lossy_compression" title="Lossy compression">lossy compression</a>.

Compressed sensing typically starts with taking a weighted linear combination of samples also called compressive measurements in a <a href="Basis_(linear_algebra)" title="Basis (linear algebra)">basis</a> different from the basis in which the signal is known to be sparse. The results found by <a href="Emmanuel_Cand%C3%A8s" title="Emmanuel Candès">Emmanuel Candès</a>, <a href="Justin_Romberg" class="mw-redirect" title="Justin Romberg">Justin Romberg</a>, <a href="Terence_Tao" title="Terence Tao">Terence Tao</a>, and <a href="David_Donoho" title="David Donoho">David Donoho</a> showed that the number of these compressive measurements can be small and still contain nearly all the useful information. Therefore, the task of converting the image back into the intended domain involves solving an underdetermined <a href="Matrix_equation" class="mw-redirect" title="Matrix equation">matrix equation</a> since the number of compressive measurements taken is smaller than the number of pixels in the full image. However, adding the constraint that the initial signal is sparse enables one to solve this underdetermined <a href="System_of_linear_equations" title="System of linear equations">system of linear equations</a>.

The least-squares solution to such problems is to minimize the <a href="L2_norm" class="mw-redirect" title="L2 norm"><math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{2}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{2}}</annotation>

</semantics>

</math><img src="./ba162c66ca85776c83557af5088cc6f8584d1912.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{2}}" loading="lazy"> norm</a>—that is, minimize the amount of energy in the system. This is usually simple mathematically (involving only a <a href="Matrix_multiplication" title="Matrix multiplication">matrix multiplication</a> by the <a href="Pseudo-inverse" class="mw-redirect" title="Pseudo-inverse">pseudo-inverse</a> of the basis sampled in). However, this leads to poor results for many practical applications, for which the unknown coefficients have nonzero energy.

To enforce the sparsity constraint when solving for the underdetermined system of linear equations, one can minimize the number of nonzero components of the solution. The function counting the number of non-zero components of a vector was called the <a href="L0_norm" class="mw-redirect" title="L0 norm"><math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{0}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{0}}</annotation>

</semantics>

</math><img src="./50d9c756a343edb4557be976651d65368e888b55.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{0}}" loading="lazy"> "norm"</a> by David Donoho.<a href="#cite_note-13">[note 1]</a>

<a href="Emmanuel_Cand%C3%A8s" title="Emmanuel Candès">Candès</a> et al. proved that for many problems it is probable that the <a href="L1_norm" class="mw-redirect" title="L1 norm"><math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy"> norm</a> is equivalent to the <a href="L0_norm" class="mw-redirect" title="L0 norm"><math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{0}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{0}}</annotation>

</semantics>

</math><img src="./50d9c756a343edb4557be976651d65368e888b55.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{0}}" loading="lazy"> norm</a>, in a technical sense: This equivalence result allows one to solve the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy"> problem, which is easier than the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{0}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{0}}</annotation>

</semantics>

</math><img src="./50d9c756a343edb4557be976651d65368e888b55.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{0}}" loading="lazy"> problem. Finding the candidate with the smallest <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{1}}</annotation>

</semantics>

</math><img src="./74c288d1089f1ec85b01b4de25c441fc792bd2d9.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{1}}" loading="lazy"> norm can be expressed relatively easily as a <a href="Linear_program" class="mw-redirect" title="Linear program">linear program</a>, for which efficient solution methods already exist.<a href="#cite_note-14">[13]</a> When measurements may contain a finite amount of noise, <a href="Basis_pursuit_denoising" title="Basis pursuit denoising">basis pursuit denoising</a> is preferred over linear programming, since it preserves sparsity in the face of noise and can be solved faster than an exact linear program.

<div class="mw-heading mw-heading3"><h3 id="Total_variation-based_CS_reconstruction">Total variation-based CS reconstruction</h3></div>

/* start https://en.wikipedia.org/ */

.mw-parser-output .hatnote{font-style:italic}.mw-parser-output div.hatnote{padding-left:1.6em;margin-bottom:0.5em}.mw-parser-output .hatnote i{font-style:normal}.mw-parser-output .hatnote+link+.hatnote{margin-top:-0.5em}@media print{body.ns-0 .mw-parser-output .hatnote{display:none!important}}

/* end https://en.wikipedia.org/ */

</style><div role="note" class="hatnote navigation-not-searchable">See also: <a href="Total-variation_denoising" class="mw-redirect" title="Total-variation denoising">Total-variation denoising</a></div>

<div class="mw-heading mw-heading4"><h4 id="Motivation_and_applications">Motivation and applications</h4></div>

<div class="mw-heading mw-heading5"><h5 id="Role_of_TV_regularization">Role of TV regularization</h5></div>

<a href="Total_variation" title="Total variation">Total variation</a> can be seen as a <a href="Non-negative" class="mw-redirect" title="Non-negative">non-negative</a> <a href="Real_number" title="Real number">real</a>-valued <a href="Functional_(mathematics)" title="Functional (mathematics)">functional</a> defined on the space of <a href="Real_number" title="Real number">real-valued</a> <a href="Function_(mathematics)" title="Function (mathematics)">functions</a> (for the case of functions of one variable) or on the space of <a href="Integrable_function" class="mw-redirect" title="Integrable function">integrable functions</a> (for the case of functions of several variables). For signals, especially, <a href="Total_variation" title="Total variation">total variation</a> refers to the integral of the absolute <a href="Gradient" title="Gradient">gradient</a> of the signal. In signal and image reconstruction, it is applied as <a href="Total_variation_regularization" class="mw-redirect" title="Total variation regularization">total variation regularization</a> where the underlying principle is that signals with excessive details have high total variation and that removing these details, while retaining important information such as edges, would reduce the total variation of the signal and make the signal subject closer to the original signal in the problem.

For the purpose of signal and image reconstruction, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \ell _{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \ell _{1}}</annotation>

</semantics>

</math><img src="./361ddd720474aa41cb05453e03424fb7999d3b02.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.024ex; height:2.509ex;" alt="{\displaystyle \ell _{1}}" loading="lazy"> minimization models are used. Other approaches also include the least-squares as has been discussed before in this article. These methods are extremely slow and return a not-so-perfect reconstruction of the signal. The current CS Regularization models attempt to address this problem by incorporating sparsity priors of the original image, one of which is the total variation (TV). Conventional TV approaches are designed to give piece-wise constant solutions. Some of these include (as discussed ahead) – constrained <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\textstyle \ell _{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\textstyle \ell _{1}}</annotation>

</semantics>

</math><img src="./4b292169108ce5c66ec29cf661510f608eec2bca.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.024ex; height:2.509ex;" alt="{\textstyle \ell _{1}}" loading="lazy">-minimization which uses an iterative scheme. This method, though fast, subsequently leads to over-smoothing of edges resulting in blurred image edges.<a href="#cite_note-EPTV-15">[14]</a> TV methods with iterative re-weighting have been implemented to reduce the influence of large gradient value magnitudes in the images. This has been used in <a href="Tomography" title="Tomography">computed tomography</a> (CT) reconstruction as a method known as edge-preserving total variation. However, as gradient magnitudes are used for estimation of relative penalty weights between the data fidelity and regularization terms, this method is not robust to noise and artifacts and accurate enough for CS image/signal reconstruction and, therefore, fails to preserve smaller structures.

Recent progress on this problem involves using an iteratively directional TV refinement for CS reconstruction.<a href="#cite_note-Orientation_and_directional_refinement-16">[15]</a> This method would have 2 stages: the first stage would estimate and refine the initial orientation field – which is defined as a noisy point-wise initial estimate, through edge-detection, of the given image. In the second stage, the CS reconstruction model is presented by utilizing directional TV regularizer. More details about these TV-based approaches – iteratively reweighted l1 minimization, edge-preserving TV and iterative model using directional orientation field and TV- are provided below.

<div class="mw-heading mw-heading4"><h4 id="Existing_approaches">Existing approaches</h4></div>

<div class="mw-heading mw-heading5"><h5 id="Iteratively_reweighted_ℓ1_minimization">Iteratively reweighted ℓ1 minimization</h5></div>

In the CS reconstruction models using constrained <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \ell _{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \ell _{1}}</annotation>

</semantics>

</math><img src="./361ddd720474aa41cb05453e03424fb7999d3b02.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.024ex; height:2.509ex;" alt="{\displaystyle \ell _{1}}" loading="lazy"> minimization,<a href="#cite_note-Original_source_for_IRLS-17">[16]</a> larger coefficients are penalized heavily in the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \ell _{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \ell _{1}}</annotation>

</semantics>

</math><img src="./361ddd720474aa41cb05453e03424fb7999d3b02.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.024ex; height:2.509ex;" alt="{\displaystyle \ell _{1}}" loading="lazy"> norm. It was proposed to have a weighted formulation of <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \ell _{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \ell _{1}}</annotation>

</semantics>

</math><img src="./361ddd720474aa41cb05453e03424fb7999d3b02.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.024ex; height:2.509ex;" alt="{\displaystyle \ell _{1}}" loading="lazy"> minimization designed to more democratically penalize nonzero coefficients. An iterative algorithm is used for constructing the appropriate weights.<a href="#cite_note-Iteration-18">[17]</a> Each iteration requires solving one <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \ell _{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \ell _{1}}</annotation>

</semantics>

</math><img src="./361ddd720474aa41cb05453e03424fb7999d3b02.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.024ex; height:2.509ex;" alt="{\displaystyle \ell _{1}}" loading="lazy"> minimization problem by finding the local minimum of a concave penalty function that more closely resembles the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \ell _{0}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \ell _{0}}</annotation>

</semantics>

</math><img src="./d18f7cb79dd41b63d6aca9ec6b957c225a0aea81.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.024ex; height:2.509ex;" alt="{\displaystyle \ell _{0}}" loading="lazy"> norm. An additional parameter, usually to avoid any sharp transitions in the penalty function curve, is introduced into the iterative equation to ensure stability and so that a zero estimate in one iteration does not necessarily lead to a zero estimate in the next iteration. The method essentially involves using the current solution for computing the weights to be used in the next iteration.

<div class="mw-heading mw-heading6"><h6 id="Advantages_and_disadvantages">Advantages and disadvantages</h6></div>

Early iterations may find inaccurate sample estimates, however this method will down-sample these at a later stage to give more weight to the smaller non-zero signal estimates. One of the disadvantages is the need for defining a valid starting point as a global minimum might not be obtained every time due to the concavity of the function. Another disadvantage is that this method tends to uniformly penalize the image gradient irrespective of the underlying image structures. This causes over-smoothing of edges, especially those of low contrast regions, subsequently leading to loss of low contrast information. The advantages of this method include: reduction of the sampling rate for sparse signals; reconstruction of the image while being robust to the removal of noise and other artifacts; and use of very few iterations. This can also help in recovering images with sparse gradients.

In the figure shown below, P1 refers to the first-step of the iterative reconstruction process, of the projection matrix P of the fan-beam geometry, which is constrained by the data fidelity term. This may contain noise and artifacts as no regularization is performed. The minimization of P1 is solved through the conjugate gradient least squares method. P2 refers to the second step of the iterative reconstruction process wherein it utilizes the edge-preserving total variation regularization term to remove noise and artifacts, and thus improve the quality of the reconstructed image/signal. The minimization of P2 is done through a simple gradient descent method. Convergence is determined by testing, after each iteration, for image positivity, by checking if <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle f^{k-1}=0}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle f^{k-1}=0}</annotation>

</semantics>

</math><img src="./935dae3cf38f1197069f1d3d4174b8960eb22cd5.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:8.771ex; height:3.009ex;" alt="{\displaystyle f^{k-1}=0}" loading="lazy"> for the case when <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle f^{k-1}<0}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle f^{k-1}<0}</annotation>

</semantics>

</math><img src="./c14953b7aaeaa3895500f63c1b70952252cb5c04.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:8.771ex; height:3.009ex;" alt="{\displaystyle f^{k-1}<0}" loading="lazy"> (Note that <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle f}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle f}</annotation>

</semantics>

</math><img src="./132e57acb643253e7810ee9702d9581f159a1c61.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:1.279ex; height:2.509ex;" alt="{\displaystyle f}" loading="lazy"> refers to the different x-ray linear attenuation coefficients at different voxels of the patient image).

<div class="mw-heading mw-heading5"><h5 id="Edge-preserving_total_variation_(TV)-based_compressed_sensing">Edge-preserving total variation (TV)-based compressed sensing</h5></div>

This is an iterative CT reconstruction algorithm with edge-preserving TV regularization to reconstruct CT images from highly undersampled data obtained at low dose CT through low current levels (milliampere). In order to reduce the imaging dose, one of the approaches used is to reduce the number of x-ray projections acquired by the scanner detectors. However, this insufficient projection data which is used to reconstruct the CT image can cause streaking artifacts. Furthermore, using these insufficient projections in standard TV algorithms end up making the problem under-determined and thus leading to infinitely many possible solutions. In this method, an additional penalty weighted function is assigned to the original TV norm. This allows for easier detection of sharp discontinuities in intensity in the images and thereby adapt the weight to store the recovered edge information during the process of signal/image reconstruction. The parameter <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \sigma }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \sigma }</annotation>

</semantics>

</math><img src="./59f59b7c3e6fdb1d0365a494b81fb9a696138c36.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.33ex; height:1.676ex;" alt="{\displaystyle \sigma }" loading="lazy"> controls the amount of smoothing applied to the pixels at the edges to differentiate them from the non-edge pixels. The value of <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \sigma }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \sigma }</annotation>

</semantics>

</math><img src="./59f59b7c3e6fdb1d0365a494b81fb9a696138c36.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.33ex; height:1.676ex;" alt="{\displaystyle \sigma }" loading="lazy"> is changed adaptively based on the values of the histogram of the gradient magnitude so that a certain percentage of pixels have gradient values larger than <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \sigma }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \sigma }</annotation>

</semantics>

</math><img src="./59f59b7c3e6fdb1d0365a494b81fb9a696138c36.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.33ex; height:1.676ex;" alt="{\displaystyle \sigma }" loading="lazy">. The edge-preserving total variation term, thus, becomes sparser and this speeds up the implementation. A two-step iteration process known as forward–backward splitting algorithm is used.<a href="#cite_note-Forward-Backward-19">[18]</a> The optimization problem is split into two sub-problems which are then solved with the conjugate gradient least squares method<a href="#cite_note-CGLS-20">[19]</a> and the simple gradient descent method respectively. The method is stopped when the desired convergence has been achieved or if the maximum number of iterations is reached.<a href="#cite_note-EPTV-15">[14]</a>

<div class="mw-heading mw-heading5"><h5 id="Advantages_and_disadvantages_2">Advantages and disadvantages</h5></div>

Some of the disadvantages of this method are the absence of smaller structures in the reconstructed image and degradation of image resolution. This edge preserving TV algorithm, however, requires fewer iterations than the conventional TV algorithm.<a href="#cite_note-EPTV-15">[14]</a> Analyzing the horizontal and vertical intensity profiles of the reconstructed images, it can be seen that there are sharp jumps at edge points and negligible, minor fluctuation at non-edge points. Thus, this method leads to low relative error and higher correlation as compared to the TV method. It also effectively suppresses and removes any form of image noise and image artifacts such as streaking.

<div class="mw-heading mw-heading5"><h5 id="Iterative_model_using_a_directional_orientation_field_and_directional_total_variation">Iterative model using a directional orientation field and directional total variation</h5></div>

To prevent over-smoothing of edges and texture details and to obtain a reconstructed CS image which is accurate and robust to noise and artifacts, this method is used. First, an initial estimate of the noisy point-wise orientation field of the image <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle I}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle I}</annotation>

</semantics>

</math><img src="./535ea7fc4134a31cbe2251d9d3511374bc41be9f.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.172ex; height:2.176ex;" alt="{\displaystyle I}" loading="lazy">, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle {\hat {d}}}">

<mover>

</mover>

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle {\hat {d}}}</annotation>

</semantics>

</math><img src="./eb237c5d0be676770778f2951db7f7f1e322dbb2.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.582ex; height:2.843ex;" alt="{\displaystyle {\hat {d}}}" loading="lazy">, is obtained. This noisy orientation field is defined so that it can be refined at a later stage to reduce the noise influences in orientation field estimation. A coarse orientation field estimation is then introduced based on structure tensor, which is formulated as:<a href="#cite_note-Structure_tensor-21">[20]</a> <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle J_{\rho }(\nabla I_{\sigma })=G_{\rho }*(\nabla I_{\sigma }\otimes \nabla I_{\sigma })={\begin{pmatrix}J_{11}&J_{12}\\J_{12}&J_{22}\end{pmatrix}}}">

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<mrow>

<mtr>

<mtd>

<msub>

</mrow>

</msub>

</mtd>

<mtd>

<msub>

</mrow>

</msub>

</mtd>

</mtr>

<mtr>

<mtd>

<msub>

</mrow>

</msub>

</mtd>

<mtd>

<msub>

</mrow>

</msub>

</mtd>

</mtr>

</mtable>

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle J_{\rho }(\nabla I_{\sigma })=G_{\rho }*(\nabla I_{\sigma }\otimes \nabla I_{\sigma })={\begin{pmatrix}J_{11}&J_{12}\\J_{12}&J_{22}\end{pmatrix}}}</annotation>

</semantics>

</math><img src="./711ec01619e151a96ebeb6dd3ca2f77059ddc2bd.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -2.505ex; width:45.354ex; height:6.176ex;" alt="{\displaystyle J_{\rho }(\nabla I_{\sigma })=G_{\rho }*(\nabla I_{\sigma }\otimes \nabla I_{\sigma })={\begin{pmatrix}J_{11}&J_{12}\\J_{12}&J_{22}\end{pmatrix}}}" loading="lazy">. Here, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle J_{\rho }}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle J_{\rho }}</annotation>

</semantics>

</math><img src="./55a611391809d16f957308ffcc0255d1768122cc.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -1.005ex; width:2.372ex; height:2.843ex;" alt="{\displaystyle J_{\rho }}" loading="lazy"> refers to the structure tensor related with the image pixel point (i,j) having standard deviation <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \rho }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \rho }</annotation>

</semantics>

</math><img src="./1f7d439671d1289b6a816e6af7a304be40608d64.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:1.202ex; height:2.176ex;" alt="{\displaystyle \rho }" loading="lazy">. <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle G}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle G}</annotation>

</semantics>

</math><img src="./f5f3c8921a3b352de45446a6789b104458c9f90b.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.827ex; height:2.176ex;" alt="{\displaystyle G}" loading="lazy"> refers to the Gaussian kernel <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle (0,\rho ^{2})}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (0,\rho ^{2})}</annotation>

</semantics>

</math><img src="./a150098f5927fa875d96e0ded1c1a353d1d52371.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:6.262ex; height:3.176ex;" alt="{\displaystyle (0,\rho ^{2})}" loading="lazy"> with standard deviation <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \rho }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \rho }</annotation>

</semantics>

</math><img src="./1f7d439671d1289b6a816e6af7a304be40608d64.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:1.202ex; height:2.176ex;" alt="{\displaystyle \rho }" loading="lazy">. <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \sigma }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \sigma }</annotation>

</semantics>

</math><img src="./59f59b7c3e6fdb1d0365a494b81fb9a696138c36.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.33ex; height:1.676ex;" alt="{\displaystyle \sigma }" loading="lazy"> refers to the manually defined parameter for the image <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle I}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle I}</annotation>

</semantics>

</math><img src="./535ea7fc4134a31cbe2251d9d3511374bc41be9f.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.172ex; height:2.176ex;" alt="{\displaystyle I}" loading="lazy"> below which the edge detection is insensitive to noise. <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \nabla I_{\sigma }}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \nabla I_{\sigma }}</annotation>

</semantics>

</math><img src="./b57f1a422d0d50e218a163ba89462e6d9ad49cff.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:4.131ex; height:2.509ex;" alt="{\displaystyle \nabla I_{\sigma }}" loading="lazy"> refers to the gradient of the image <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle I}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle I}</annotation>

</semantics>

</math><img src="./535ea7fc4134a31cbe2251d9d3511374bc41be9f.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.172ex; height:2.176ex;" alt="{\displaystyle I}" loading="lazy"> and <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle (\nabla I_{\sigma }\otimes \nabla I_{\sigma })}">

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (\nabla I_{\sigma }\otimes \nabla I_{\sigma })}</annotation>

</semantics>

</math><img src="./c81ba01249cc8f87a5fb8db7228fdeca24ccc127.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:12.913ex; height:2.843ex;" alt="{\displaystyle (\nabla I_{\sigma }\otimes \nabla I_{\sigma })}" loading="lazy"> refers to the tensor product obtained by using this gradient.<a href="#cite_note-Orientation_and_directional_refinement-16">[15]</a>

The structure tensor obtained is convolved with a Gaussian kernel <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle G}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle G}</annotation>

</semantics>

</math><img src="./f5f3c8921a3b352de45446a6789b104458c9f90b.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.827ex; height:2.176ex;" alt="{\displaystyle G}" loading="lazy"> to improve the accuracy of the orientation estimate with <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \sigma }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \sigma }</annotation>

</semantics>

</math><img src="./59f59b7c3e6fdb1d0365a494b81fb9a696138c36.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.33ex; height:1.676ex;" alt="{\displaystyle \sigma }" loading="lazy"> being set to high values to account for the unknown noise levels. For every pixel (i,j) in the image, the structure tensor J is a symmetric and positive semi-definite matrix. Convolving all the pixels in the image with <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle G}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle G}</annotation>

</semantics>

</math><img src="./f5f3c8921a3b352de45446a6789b104458c9f90b.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.827ex; height:2.176ex;" alt="{\displaystyle G}" loading="lazy">, gives orthonormal eigen vectors ω and υ of the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle J}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle J}</annotation>

</semantics>

</math><img src="./359e4f407b49910e02c27c2f52e87a36cd74c053.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.471ex; height:2.176ex;" alt="{\displaystyle J}" loading="lazy"> matrix. ω points in the direction of the dominant orientation having the largest contrast and υ points in the direction of the structure orientation having the smallest contrast. The orientation field coarse initial estimation <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle {\hat {d}}}">

<mover>

</mover>

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle {\hat {d}}}</annotation>

</semantics>

</math><img src="./eb237c5d0be676770778f2951db7f7f1e322dbb2.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.582ex; height:2.843ex;" alt="{\displaystyle {\hat {d}}}" loading="lazy"> is defined as <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle {\hat {d}}}">

<mover>

</mover>

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle {\hat {d}}}</annotation>

</semantics>

</math><img src="./eb237c5d0be676770778f2951db7f7f1e322dbb2.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.582ex; height:2.843ex;" alt="{\displaystyle {\hat {d}}}" loading="lazy"> = υ. This estimate is accurate at strong edges. However, at weak edges or on regions with noise, its reliability decreases.

To overcome this drawback, a refined orientation model is defined in which the data term reduces the effect of noise and improves accuracy while the second penalty term with the L2-norm is a fidelity term which ensures accuracy of initial coarse estimation.

This orientation field is introduced into the directional total variation optimization model for CS reconstruction through the equation: <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \min _{\mathrm {X} }\lVert \nabla \mathrm {X} \bullet d\rVert _{1}+{\frac {\lambda }{2}}\ \lVert Y-\Phi \mathrm {X} \rVert _{2}^{2}}">

</mrow>

</munder>

</mrow>

<msub>

</mrow>

</msub>

<mfrac>

</mfrac>

</mrow>

</mrow>

</mrow>

</mrow>

</msubsup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \min _{\mathrm {X} }\lVert \nabla \mathrm {X} \bullet d\rVert _{1}+{\frac {\lambda }{2}}\ \lVert Y-\Phi \mathrm {X} \rVert _{2}^{2}}</annotation>

</semantics>

</math><img src="./c84afc8da52c92a79dc209d083e91e956c33937e.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -2.005ex; width:31.37ex; height:5.509ex;" alt="{\displaystyle \min _{\mathrm {X} }\lVert \nabla \mathrm {X} \bullet d\rVert _{1}+{\frac {\lambda }{2}}\ \lVert Y-\Phi \mathrm {X} \rVert _{2}^{2}}" loading="lazy">. <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathrm {X} }">

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathrm {X} }</annotation>

</semantics>

</math><img src="./a8d98cdc2216c7394d189ea3e09a479c826263b6.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.743ex; height:2.176ex;" alt="{\displaystyle \mathrm {X} }" loading="lazy"> is the objective signal which needs to be recovered. Y is the corresponding measurement vector, d is the iterative refined orientation field and <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \Phi }">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \Phi }</annotation>

</semantics>

</math><img src="./aed80a2011a3912b028ba32a52dfa57165455f24.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.678ex; height:2.176ex;" alt="{\displaystyle \Phi }" loading="lazy"> is the CS measurement matrix. This method undergoes a few iterations ultimately leading to convergence.<math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle {\hat {d}}}">

<mover>

</mover>

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle {\hat {d}}}</annotation>

</semantics>

</math><img src="./eb237c5d0be676770778f2951db7f7f1e322dbb2.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.582ex; height:2.843ex;" alt="{\displaystyle {\hat {d}}}" loading="lazy"> is the orientation field approximate estimation of the reconstructed image <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle X^{k-1}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle X^{k-1}}</annotation>

</semantics>

</math><img src="./ecc0413c1068707f9e27c7e5c61dab760795a306.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:5.186ex; height:2.676ex;" alt="{\displaystyle X^{k-1}}" loading="lazy"> from the previous iteration (in order to check for convergence and the subsequent optical performance, the previous iteration is used). For the two vector fields represented by <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathrm {X} }">

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathrm {X} }</annotation>

</semantics>

</math><img src="./a8d98cdc2216c7394d189ea3e09a479c826263b6.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.743ex; height:2.176ex;" alt="{\displaystyle \mathrm {X} }" loading="lazy"> and <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle d}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle d}</annotation>

</semantics>

</math><img src="./e85ff03cbe0c7341af6b982e47e9f90d235c66ab.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.216ex; height:2.176ex;" alt="{\displaystyle d}" loading="lazy">, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathrm {X} \bullet d}">

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathrm {X} \bullet d}</annotation>

</semantics>

</math><img src="./93d54e65fd11239dacb9ba79fbcbd344fb09a455.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:5.154ex; height:2.176ex;" alt="{\displaystyle \mathrm {X} \bullet d}" loading="lazy"> refers to the multiplication of respective horizontal and vertical vector elements of <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathrm {X} }">

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathrm {X} }</annotation>

</semantics>

</math><img src="./a8d98cdc2216c7394d189ea3e09a479c826263b6.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.743ex; height:2.176ex;" alt="{\displaystyle \mathrm {X} }" loading="lazy"> and <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle d}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle d}</annotation>

</semantics>

</math><img src="./e85ff03cbe0c7341af6b982e47e9f90d235c66ab.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.216ex; height:2.176ex;" alt="{\displaystyle d}" loading="lazy"> followed by their subsequent addition. These equations are reduced to a series of convex minimization problems which are then solved with a combination of <a href="Variable_splitting" title="Variable splitting">variable splitting</a> and augmented Lagrangian (FFT-based fast solver with a closed form solution) methods.<a href="#cite_note-Orientation_and_directional_refinement-16">[15]</a> It (Augmented Lagrangian) is considered equivalent to the split Bregman iteration which ensures convergence of this method. The orientation field, d is defined as being equal to <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle (d_{h},d_{v})}">

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (d_{h},d_{v})}</annotation>

</semantics>

</math><img src="./b64b06c49922d5c99b77a17280e12af286e0fa36.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:7.47ex; height:2.843ex;" alt="{\displaystyle (d_{h},d_{v})}" loading="lazy">, where <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle d_{h},d_{v}}">

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle d_{h},d_{v}}</annotation>

</semantics>

</math><img src="./f61bd5b74da507136bbb87a8e2950139a9ed3518.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:5.66ex; height:2.509ex;" alt="{\displaystyle d_{h},d_{v}}" loading="lazy"> define the horizontal and vertical estimates of <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle d}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle d}</annotation>

</semantics>

</math><img src="./e85ff03cbe0c7341af6b982e47e9f90d235c66ab.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.216ex; height:2.176ex;" alt="{\displaystyle d}" loading="lazy">.

The Augmented Lagrangian method for the orientation field, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \min _{\mathrm {X} }\lVert \nabla \mathrm {X} \bullet d\rVert _{1}+{\frac {\lambda }{2}}\ \lVert Y-\Phi \mathrm {X} \rVert _{2}^{2}}">

</mrow>

</munder>

</mrow>

<msub>

</mrow>

</msub>

<mfrac>

</mfrac>

</mrow>

</mrow>

</mrow>

</mrow>

</msubsup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \min _{\mathrm {X} }\lVert \nabla \mathrm {X} \bullet d\rVert _{1}+{\frac {\lambda }{2}}\ \lVert Y-\Phi \mathrm {X} \rVert _{2}^{2}}</annotation>

</semantics>

</math><img src="./c84afc8da52c92a79dc209d083e91e956c33937e.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -2.005ex; width:31.37ex; height:5.509ex;" alt="{\displaystyle \min _{\mathrm {X} }\lVert \nabla \mathrm {X} \bullet d\rVert _{1}+{\frac {\lambda }{2}}\ \lVert Y-\Phi \mathrm {X} \rVert _{2}^{2}}" loading="lazy">, involves initializing <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle d_{h},d_{v},H,V}">

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle d_{h},d_{v},H,V}</annotation>

</semantics>

</math><img src="./8de0f389ccbcf3b146acf9fdffc919c8a9a52abe.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:11.579ex; height:2.509ex;" alt="{\displaystyle d_{h},d_{v},H,V}" loading="lazy"> and then finding the approximate minimizer of <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L_{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L_{1}}</annotation>

</semantics>

</math><img src="./0e79dc1b001f8b923df475ed14de023cbc456013.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.637ex; height:2.509ex;" alt="{\displaystyle L_{1}}" loading="lazy"> with respect to these variables. The Lagrangian multipliers are then updated and the iterative process is stopped when convergence is achieved. For the iterative directional total variation refinement model, the augmented lagrangian method involves initializing <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathrm {X} ,P,Q,\lambda _{P},\lambda _{Q}}">

</mrow>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathrm {X} ,P,Q,\lambda _{P},\lambda _{Q}}</annotation>

</semantics>

</math><img src="./78eba797338e19fd9eb3f96725df3191ebf88d40.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -1.005ex; width:15.172ex; height:2.843ex;" alt="{\displaystyle \mathrm {X} ,P,Q,\lambda _{P},\lambda _{Q}}" loading="lazy">.<a href="#cite_note-TV-22">[21]</a>

Here, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle H,V,P,Q}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle H,V,P,Q}</annotation>

</semantics>

</math><img src="./925cea510d1e9497d80a2b9e20a6e90751251ee4.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:10.536ex; height:2.509ex;" alt="{\displaystyle H,V,P,Q}" loading="lazy"> are newly introduced variables where <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle H}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle H}</annotation>

</semantics>

</math><img src="./75a9edddcca2f782014371f75dca39d7e13a9c1b.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.064ex; height:2.176ex;" alt="{\displaystyle H}" loading="lazy"> = <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \nabla d_{h}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \nabla d_{h}}</annotation>

</semantics>

</math><img src="./25deb69dca59f312cdd5ec5e03c3fc9e0543867d.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:4.324ex; height:2.509ex;" alt="{\displaystyle \nabla d_{h}}" loading="lazy">, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle V}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle V}</annotation>

</semantics>

</math><img src="./af0f6064540e84211d0ffe4dac72098adfa52845.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.787ex; height:2.176ex;" alt="{\displaystyle V}" loading="lazy"> = <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \nabla d_{v}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \nabla d_{v}}</annotation>

</semantics>

</math><img src="./548291cd7fd3d9eb4e0a2b5a20efd13874057122.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:4.174ex; height:2.509ex;" alt="{\displaystyle \nabla d_{v}}" loading="lazy">, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle P}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle P}</annotation>

</semantics>

</math><img src="./b4dc73bf40314945ff376bd363916a738548d40a.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:1.745ex; height:2.176ex;" alt="{\displaystyle P}" loading="lazy"> = <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \nabla \mathrm {X} }">

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \nabla \mathrm {X} }</annotation>

</semantics>

</math><img src="./946e06596049d356220e90c090c13d0932da3a35.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:3.679ex; height:2.176ex;" alt="{\displaystyle \nabla \mathrm {X} }" loading="lazy">, and <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle Q}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle Q}</annotation>

</semantics>

</math><img src="./8752c7023b4b3286800fe3238271bbca681219ed.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:1.838ex; height:2.509ex;" alt="{\displaystyle Q}" loading="lazy"> = <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle P\bullet d}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle P\bullet d}</annotation>

</semantics>

</math><img src="./4c38df68bcc0d259fd708cbe23797652cc2402f0.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:5.156ex; height:2.176ex;" alt="{\displaystyle P\bullet d}" loading="lazy">. <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \lambda _{H},\lambda _{V},\lambda _{P},\lambda _{Q}}">

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \lambda _{H},\lambda _{V},\lambda _{P},\lambda _{Q}}</annotation>

</semantics>

</math><img src="./7f174ee7e44f9d570c8e7b2b17690b2bb92be1dd.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -1.005ex; width:14.709ex; height:2.843ex;" alt="{\displaystyle \lambda _{H},\lambda _{V},\lambda _{P},\lambda _{Q}}" loading="lazy"> are the Lagrangian multipliers for <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle H,V,P,Q}">

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle H,V,P,Q}</annotation>

</semantics>

</math><img src="./925cea510d1e9497d80a2b9e20a6e90751251ee4.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:10.536ex; height:2.509ex;" alt="{\displaystyle H,V,P,Q}" loading="lazy">. For each iteration, the approximate minimizer of <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L_{2}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L_{2}}</annotation>

</semantics>

</math><img src="./c6a952cfe42c86b7741f55a817da0e251793a358.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:2.637ex; height:2.509ex;" alt="{\displaystyle L_{2}}" loading="lazy"> with respect to variables (<math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \mathrm {X} ,P,Q}">

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \mathrm {X} ,P,Q}</annotation>

</semantics>

</math><img src="./516fcd6db1dd06abd50507324444877365475f86.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:7.395ex; height:2.509ex;" alt="{\displaystyle \mathrm {X} ,P,Q}" loading="lazy">) is calculated. And as in the field refinement model, the lagrangian multipliers are updated and the iterative process is stopped when convergence is achieved.

For the orientation field refinement model, the Lagrangian multipliers are updated in the iterative process as follows:

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (\lambda _{H})^{k}=(\lambda _{H})^{k-1}+\gamma _{H}(H^{k}-\nabla (d_{h})^{k})}</annotation>

</semantics>

</math><img src="./82bbb9bd8fb56ab4b277d83683005bf8c191c2cc.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:37.888ex; height:3.176ex;" alt="{\displaystyle (\lambda _{H})^{k}=(\lambda _{H})^{k-1}+\gamma _{H}(H^{k}-\nabla (d_{h})^{k})}" loading="lazy"></dd></dl>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (\lambda _{V})^{k}=(\lambda _{V})^{k-1}+\gamma _{V}(V^{k}-\nabla (d_{v})^{k})}</annotation>

</semantics>

</math><img src="./5754d34f442fa3b63d56de2d318e328d49d612bc.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:36.966ex; height:3.176ex;" alt="{\displaystyle (\lambda _{V})^{k}=(\lambda _{V})^{k-1}+\gamma _{V}(V^{k}-\nabla (d_{v})^{k})}" loading="lazy"></dd></dl>

For the iterative directional total variation refinement model, the Lagrangian multipliers are updated as follows:

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

</mrow>

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (\lambda _{P})^{k}=(\lambda _{P})^{k-1}+\gamma _{P}P^{k}-\nabla (\mathrm {X} )^{k})}</annotation>

</semantics>

</math><img src="./d948121ab1b7647ae49f01fd87e0084016d1c01b.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.838ex; width:35.382ex; height:3.176ex;" alt="{\displaystyle (\lambda _{P})^{k}=(\lambda _{P})^{k-1}+\gamma _{P}P^{k}-\nabla (\mathrm {X} )^{k})}" loading="lazy"></dd></dl>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

<msup>

</mrow>

</msup>

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (\lambda _{Q})^{k}=(\lambda _{Q})^{k-1}+\gamma _{Q}(Q^{k}-P^{k}\bullet d)}</annotation>

</semantics>

</math><img src="./ae62071a2b7a357eb84afa12f1f1fcecfb342d62.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -1.005ex; width:36.244ex; height:3.343ex;" alt="{\displaystyle (\lambda _{Q})^{k}=(\lambda _{Q})^{k-1}+\gamma _{Q}(Q^{k}-P^{k}\bullet d)}" loading="lazy"></dd></dl>

Here, <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle \gamma _{H},\gamma _{V},\gamma _{P},\gamma _{Q}}">

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle \gamma _{H},\gamma _{V},\gamma _{P},\gamma _{Q}}</annotation>

</semantics>

</math><img src="./ec7f020a0c9c8c11582b4cc3bcade561473b66de.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -1.005ex; width:14.105ex; height:2.343ex;" alt="{\displaystyle \gamma _{H},\gamma _{V},\gamma _{P},\gamma _{Q}}" loading="lazy"> are positive constants.

<div class="mw-heading mw-heading5"><h5 id="Advantages_and_disadvantages_3">Advantages and disadvantages</h5></div>

Based on <a href="Peak_signal-to-noise_ratio" title="Peak signal-to-noise ratio">peak signal-to-noise ratio</a> (PSNR) and <a href="Structural_similarity" class="mw-redirect" title="Structural similarity">structural similarity</a> index (SSIM) metrics and known ground-truth images for testing performance, it is concluded that iterative directional total variation has a better reconstructed performance than the non-iterative methods in preserving edge and texture areas. The orientation field refinement model plays a major role in this improvement in performance as it increases the number of directionless pixels in the flat area while enhancing the orientation field consistency in the regions with edges.

<div class="mw-heading mw-heading2"><h2 id="Applications">Applications</h2></div>

The field of compressive sensing is related to several topics in signal processing and computational mathematics, such as <a href="Underdetermined_system" title="Underdetermined system">underdetermined linear systems</a>, <a href="Group_testing" title="Group testing">group testing</a>, heavy hitters, <a href="Sparse_coding" class="mw-redirect" title="Sparse coding">sparse coding</a>, <a href="Multiplexing" title="Multiplexing">multiplexing</a>, sparse sampling, and finite rate of innovation. Its broad scope and generality has enabled several innovative CS-enhanced approaches in signal processing and compression, solution of inverse problems, design of radiating systems, radar and through-the-wall imaging, and antenna characterization.<a href="#cite_note-23">[22]</a> Imaging techniques having a strong affinity with compressive sensing include <a href="Coded_aperture" title="Coded aperture">coded aperture</a> and <a href="Computational_photography" title="Computational photography">computational photography</a>.

Conventional CS reconstruction uses sparse signals (usually sampled at a rate less than the Nyquist sampling rate) for reconstruction through constrained <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle l_{1}}">

<msub>

</mrow>

</msub>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle l_{1}}</annotation>

</semantics>

</math><img src="./29b25eeca673386d676f79dce674fe93040693eb.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.671ex; width:1.748ex; height:2.509ex;" alt="{\displaystyle l_{1}}" loading="lazy"> minimization. One of the earliest applications of such an approach was in reflection seismology which used sparse reflected signals from band-limited data for tracking changes between sub-surface layers.<a href="#cite_note-Seismic_sparse_signals-24">[23]</a> When the LASSO model came into prominence in the 1990s as a statistical method for selection of sparse models,<a href="#cite_note-LASSO-25">[24]</a> this method was further used in computational harmonic analysis for sparse signal representation from over-complete dictionaries. Some of the other applications include incoherent sampling of radar pulses. The work by Boyd et al.<a href="#cite_note-Original_source_for_IRLS-17">[16]</a> has applied the LASSO model- for selection of sparse models- towards analog to digital converters (the current ones use a sampling rate higher than the Nyquist rate along with the quantized Shannon representation). This would involve a parallel architecture in which the polarity of the analog signal changes at a high rate followed by digitizing the integral at the end of each time-interval to obtain the converted digital signal.

<div class="mw-heading mw-heading3"><h3 id="Photography">Photography</h3></div>

Compressed sensing has been used in an experimental mobile phone camera sensor. The approach allows a reduction in image acquisition energy per image by as much as a factor of 15 at the cost of complex decompression algorithms; the computation may require an off-device implementation.<a href="#cite_note-26">[25]</a>

Compressed sensing is used in single-pixel cameras from <a href="Rice_University" title="Rice University">Rice University</a>.<a href="#cite_note-cscamera-27">[26]</a> <a href="Bell_Labs" title="Bell Labs">Bell Labs</a> employed the technique in a lensless single-pixel camera that takes stills using repeated snapshots of randomly chosen apertures from a grid. Image quality improves with the number of snapshots, and generally requires a small fraction of the data of conventional imaging, while eliminating lens/focus-related aberrations.<a href="#cite_note-28">[27]</a><a href="#cite_note-29">[28]</a>

<div class="mw-heading mw-heading3"><h3 id="Holography">Holography</h3></div>

Compressed sensing can be used to improve image reconstruction in <a href="Holography" title="Holography">holography</a> by increasing the number of <a href="Voxel" title="Voxel">voxels</a> one can infer from a single hologram.<a href="#cite_note-30">[29]</a><a href="#cite_note-31">[30]</a><a href="#cite_note-32">[31]</a> It is also used for image retrieval from undersampled measurements in optical<a href="#cite_note-33">[32]</a><a href="#cite_note-34">[33]</a> and millimeter-wave<a href="#cite_note-35">[34]</a> holography.

<div class="mw-heading mw-heading3"><h3 id="Facial_recognition">Facial recognition</h3></div>

Compressed sensing has been used in <a href="Facial_recognition_system" title="Facial recognition system">facial recognition</a> applications.<a href="#cite_note-36">[35]</a>

<div class="mw-heading mw-heading3"><h3 id="Magnetic_resonance_imaging">Magnetic resonance imaging</h3></div>

Compressed sensing has been used<a href="#cite_note-dx.doi.org-37">[36]</a><a href="#cite_note-Compressed_Sensing_MRI_2008-38">[37]</a> to shorten <a href="Magnetic_resonance_imaging" title="Magnetic resonance imaging">magnetic resonance imaging</a> scanning sessions on conventional hardware.<a href="#cite_note-39">[38]</a> Reconstruction methods include

<li>FISTA</li>

<li>SISTA</li>

<li>ePRESS<a href="#cite_note-40">[39]</a></li>

<li>EWISTA<a href="#cite_note-Zhang_2015-41">[40]</a></li>

<li>EWISTARS<a href="#cite_note-42">[41]</a> etc.</li></ul>

Compressed sensing addresses the issue of high scan time by enabling faster acquisition by measuring fewer Fourier coefficients. This produces a high-quality image with relatively lower scan time. Another application (also discussed ahead) is for CT reconstruction with fewer X-ray projections. Compressed sensing, in this case, removes the high spatial gradient parts – mainly, image noise and artifacts. This holds potential as one can obtain high-resolution CT images at low radiation doses (through lower current-mA settings).<a href="#cite_note-MRI-43">[42]</a>

<div class="mw-heading mw-heading3"><h3 id="Network_tomography">Network tomography</h3></div>

Compressed sensing has shown outstanding results in the application of <a href="Network_tomography" title="Network tomography">network tomography</a> to <a href="Network_management" title="Network management">network management</a>. <a href="Network_delay" title="Network delay">Network delay</a> estimation and <a href="Network_congestion" title="Network congestion">network congestion</a> detection can both be modeled as underdetermined <a href="System_of_linear_equations" title="System of linear equations">systems of linear equations</a> where the coefficient matrix is the network routing matrix. Moreover, in the <a href="Internet" title="Internet">Internet</a>, network routing matrices usually satisfy the criterion for using compressed sensing.<a href="#cite_note-44">[43]</a>

<div class="mw-heading mw-heading3"><h3 id="Shortwave-infrared_cameras">Shortwave-infrared cameras</h3></div>

In 2013 one company announced shortwave-infrared cameras which utilize compressed sensing.<a href="#cite_note-45">[44]</a> These cameras have light sensitivity from 0.9 <a href="%CE%9Cm" class="mw-redirect" title="Μm">μm</a> to 1.7 μm, wavelengths invisible to the human eye.

<div class="mw-heading mw-heading3"><h3 id="Aperture_synthesis_astronomy">Aperture synthesis astronomy</h3></div>

In <a href="Radio_astronomy" title="Radio astronomy">radio astronomy</a> and optical <a href="Astronomical_interferometry" class="mw-redirect" title="Astronomical interferometry">astronomical interferometry</a>, full coverage of the Fourier plane is usually absent and phase information is not obtained in most hardware configurations. In order to obtain <a href="Aperture_synthesis" title="Aperture synthesis">aperture synthesis</a> images, various compressed sensing algorithms are employed.<a href="#cite_note-46">[45]</a> The <a href="CLEAN_(algorithm)" title="CLEAN (algorithm)">Högbom CLEAN algorithm</a> has been in use since 1974 for the reconstruction of images obtained from radio interferometers, which is similar to the <a href="Matching_pursuit" title="Matching pursuit">matching pursuit</a> algorithm mentioned above.

<div class="mw-heading mw-heading3"><h3 id="Transmission_electron_microscopy">Transmission electron microscopy</h3></div>

Compressed sensing combined with a moving aperture has been used to increase the acquisition rate of images in a <a href="Transmission_electron_microscopy" title="Transmission electron microscopy">transmission electron microscope</a>.<a href="#cite_note-47">[46]</a> In <a href="Scanning_transmission_electron_microscopy" title="Scanning transmission electron microscopy">scanning mode</a>, compressive sensing combined with random scanning of the electron beam has enabled both faster acquisition and less electron dose, which allows for imaging of electron beam sensitive materials.<a href="#cite_note-48">[47]</a>

<ul><li><a href="Compressed_sensing_in_speech_signals" title="Compressed sensing in speech signals">Compressed sensing in speech signals</a></li>

<li><a href="Low-density_parity-check_code" title="Low-density parity-check code">Low-density parity-check code</a></li>

<li><a href="Noiselet" title="Noiselet">Noiselet</a></li>

<li><a href="Sparse_approximation" title="Sparse approximation">Sparse approximation</a></li>

<li><a href="Sparse_coding" class="mw-redirect" title="Sparse coding">Sparse coding</a></li>

<li><a href="Verification-based_message-passing_algorithms_in_compressed_sensing" title="Verification-based message-passing algorithms in compressed sensing">Verification-based message-passing algorithms in compressed sensing</a></li></ul>

<div class="mw-heading mw-heading2"><h2 id="Notes">Notes</h2></div>

/* start https://en.wikipedia.org/ */

.mw-parser-output .reflist{margin-bottom:0.5em;list-style-type:decimal}@media screen{.mw-parser-output .reflist{font-size:90%}}.mw-parser-output .reflist .references{font-size:100%;margin-bottom:0;list-style-type:inherit}.mw-parser-output .reflist-columns-2{column-width:30em}.mw-parser-output .reflist-columns-3{column-width:25em}.mw-parser-output .reflist-columns{margin-top:0.3em}.mw-parser-output .reflist-columns ol{margin-top:0}.mw-parser-output .reflist-columns li{page-break-inside:avoid;break-inside:avoid-column}.mw-parser-output .reflist-upper-alpha{list-style-type:upper-alpha}.mw-parser-output .reflist-upper-roman{list-style-type:upper-roman}.mw-parser-output .reflist-lower-alpha{list-style-type:lower-alpha}.mw-parser-output .reflist-lower-greek{list-style-type:lower-greek}.mw-parser-output .reflist-lower-roman{list-style-type:lower-roman}

/* end https://en.wikipedia.org/ */

</style><div class="reflist">

<li id="cite_note-13"><a href="#cite_ref-13">^</a> The quotation marks served two warnings. First, the number-of-nonzeros <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{0}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{0}}</annotation>

</semantics>

</math><img src="./50d9c756a343edb4557be976651d65368e888b55.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{0}}" loading="lazy">-"norm" is not a proper <a href="F-space" title="F-space">F-norm</a>, because it is not continuous in its scalar argument: nnzs(αx) is constant as α approaches zero. Unfortunately, authors now neglect the quotation marks and <a href="Abuse_of_terminology" class="mw-redirect" title="Abuse of terminology">abused terminology</a>—clashing with the established use of the <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle L^{0}}">

<msup>

</mrow>

</msup>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle L^{0}}</annotation>

</semantics>

</math><img src="./50d9c756a343edb4557be976651d65368e888b55.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -0.338ex; width:2.637ex; height:2.676ex;" alt="{\displaystyle L^{0}}" loading="lazy"> norm for the space of measurable functions (equipped with an appropriate metric) or for the <a href="F-space" title="F-space">space</a> of sequences with <a href="F-space" title="F-space">F–norm</a> <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="{\displaystyle (x_{n})\mapsto \sum _{n}{2^{-n}x_{n}/(1+x_{n})}}">

<msub>

</mrow>

</msub>

</mrow>

</munder>

<msup>

</mrow>

</msup>

<msub>

</mrow>

</msub>

</mrow>

<msub>

</mrow>

</msub>

</mrow>

</mstyle>

</mrow>

<annotation encoding="application/x-tex">{\displaystyle (x_{n})\mapsto \sum _{n}{2^{-n}x_{n}/(1+x_{n})}}</annotation>

</semantics>

</math><img src="./054e78aa8e7679a1979bf5f80924c51e907fd943.svg" class="mwe-math-fallback-image-inline mw-invert skin-invert" aria-hidden="true" style="vertical-align: -3.005ex; width:27.444ex; height:5.509ex;" alt="{\displaystyle (x_{n})\mapsto \sum _{n}{2^{-n}x_{n}/(1+x_{n})}}" loading="lazy">.<a href="#cite_note-12">[12]</a>

</li>

</ol></div></div>

<div class="mw-heading mw-heading2"><h2 id="References">References</h2></div>

/* start https://en.wikipedia.org/ */

.mw-parser-output cite.citation{font-style:inherit;word-wrap:break-word}.mw-parser-output .citation q{quotes:"\"""\"""'""'"}.mw-parser-output .citation:target{background-color:rgba(0,127,255,0.133)}.mw-parser-output .id-lock-free.id-lock-free a{background:url("./mw/Lock-green.svg")right 0.1em center/9px no-repeat}.mw-parser-output .id-lock-limited.id-lock-limited a,.mw-parser-output .id-lock-registration.id-lock-registration a{background:url("./mw/Lock-gray-alt-2.svg")right 0.1em center/9px no-repeat}.mw-parser-output .id-lock-subscription.id-lock-subscription a{background:url("./mw/Lock-red-alt-2.svg")right 0.1em center/9px no-repeat}.mw-parser-output .cs1-ws-icon a{background:url("./mw/Wikisource-logo.svg")right 0.1em center/12px no-repeat}body:not(.skin-timeless):not(.skin-minerva) .mw-parser-output .id-lock-free a,body:not(.skin-timeless):not(.skin-minerva) .mw-parser-output .id-lock-limited a,body:not(.skin-timeless):not(.skin-minerva) .mw-parser-output .id-lock-registration a,body:not(.skin-timeless):not(.skin-minerva) .mw-parser-output .id-lock-subscription a,body:not(.skin-timeless):not(.skin-minerva) .mw-parser-output .cs1-ws-icon a{background-size:contain;padding:0 1em 0 0}.mw-parser-output .cs1-code{color:inherit;background:inherit;border:none;padding:inherit}.mw-parser-output .cs1-hidden-error{display:none;color:var(--color-error,#d33)}.mw-parser-output .cs1-visible-error{color:var(--color-error,#d33)}.mw-parser-output .cs1-maint{display:none;color:#085;margin-left:0.3em}.mw-parser-output .cs1-kern-left{padding-left:0.2em}.mw-parser-output .cs1-kern-right{padding-right:0.2em}.mw-parser-output .citation .mw-selflink{font-weight:inherit}@media screen{.mw-parser-output .cs1-format{font-size:95%}html.skin-theme-clientpref-night .mw-parser-output .cs1-maint{color:#18911f}}@media screen and (prefers-color-scheme:dark){html.skin-theme-clientpref-os .mw-parser-output .cs1-maint{color:#18911f}}

/* end https://en.wikipedia.org/ */

</style><cite id="CITEREFDonoho2006" class="citation journal cs1">Donoho, David L. (2006). "For most large underdetermined systems of linear equations the minimal 1-norm solution is also the sparsest solution". Communications on Pure and Applied Mathematics. 59 (6): 797–829. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1002%2Fcpa.20132">10.1002/cpa.20132</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:8510060">8510060</a>.</cite>

</li>

<li id="cite_note-2"><a href="#cite_ref-2">^</a> M. Davenport, <a rel="nofollow" class="external text" href="http://www.brainshark.com/brainshark/brainshark.net/portal/title.aspx?pid=zCdz10BfTRz0z0">"The Fundamentals of Compressive Sensing"</a>, SigView, April 12, 2013.

</li>

<li id="cite_note-3"><a href="#cite_ref-3">^</a> Candès, E.J., & Plan, Y. (2010). A Probabilistic and RIPless Theory of Compressed Sensing. IEEE Transactions on Information Theory, 57, 7235–7254.

</li>

<li id="cite_note-4"><a href="#cite_ref-4">^</a> <cite id="CITEREFCandèsRombergTao2006" class="citation journal cs1">Candès, Emmanuel J.; Romberg, Justin K.; Tao, Terence (2006). <a rel="nofollow" class="external text" href="https://web.archive.org/web/20120311192231/http://www-stat.stanford.edu/~candes/papers/StableRecovery.pdf">"Stable signal recovery from incomplete and inaccurate measurements"</a> (PDF). Communications on Pure and Applied Mathematics. 59 (8): 1207–1223. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/math/0503066">math/0503066</a>. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2005math......3066C">2005math......3066C</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1002%2Fcpa.20124">10.1002/cpa.20124</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:119159284">119159284</a>. Archived from <a rel="nofollow" class="external text" href="http://www-stat.stanford.edu/~candes/papers/StableRecovery.pdf">the original</a> (PDF) on 2012-03-11. Retrieved 2011-02-10.</cite>

</li>

<li id="cite_note-Donoho-5"><a href="#cite_ref-Donoho_5-0">^</a> <cite id="CITEREFDonoho2006" class="citation journal cs1">Donoho, D.L. (2006). "Compressed sensing". IEEE Transactions on Information Theory. 52 (4): 1289–1306. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2FTIT.2006.871582">10.1109/TIT.2006.871582</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:206737254">206737254</a>.</cite>

</li>

<li id="cite_note-6"><a href="#cite_ref-6">^</a> <a rel="nofollow" class="external text" href="http://2.bp.blogspot.com/_0ZCyAOBrUtA/TTwqLEeLvdI/AAAAAAAAEXI/7S0_SnWoC0E/s1600/l1-minimization.JPG">List of L1 regularization ideas</a> from Vivek Goyal, Alyson Fletcher, Sundeep Rangan, <a rel="nofollow" class="external text" href="http://www.math.uiuc.edu/%7Elaugesen/imaha10/goyal_talk.pdf">The Optimistic Bayesian: Replica Method Analysis of Compressed Sensing</a>

</li>

<li id="cite_note-7"><a href="#cite_ref-7">^</a> <cite id="CITEREFHayes2009" class="citation journal cs1">Hayes, Brian (2009). "The Best Bits". American Scientist. 97 (4): 276. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1511%2F2009.79.276">10.1511/2009.79.276</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:349102">349102</a>.</cite>

</li>

<li id="cite_note-8"><a href="#cite_ref-8">^</a> <cite id="CITEREFTibshirani1996" class="citation journal cs1">Tibshirani, Robert (1996). <a rel="nofollow" class="external text" href="http://www-stat.stanford.edu/~tibs/lasso.html">"Regression shrinkage and selection via the lasso"</a>. <a href="Journal_of_the_Royal_Statistical_Society%2C_Series_B" class="mw-redirect" title="Journal of the Royal Statistical Society, Series B">Journal of the Royal Statistical Society, Series B</a>. 58 (1): 267–288. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1111%2Fj.2517-6161.1996.tb02080.x">10.1111/j.2517-6161.1996.tb02080.x</a>.</cite>

</li>

<li id="cite_note-9"><a href="#cite_ref-9">^</a> "Atomic decomposition by basis pursuit", by Scott Shaobing Chen, David L. Donoho, Michael, A. Saunders. SIAM Journal on Scientific Computing

</li>

<li id="cite_note-10"><a href="#cite_ref-10">^</a> <cite id="CITEREFCandèsRombergTao2006" class="citation journal cs1">Candès, Emmanuel J.; Romberg, Justin K.; Tao, Terence (2006). <a rel="nofollow" class="external text" href="http://www-stat.stanford.edu/~candes/papers/ExactRecovery.pdf">"Robust Uncertainty Principles: Exact Signal Reconstruction from Highly Incomplete Fourier Information"</a> (PDF). IEEE Trans. Inf. Theory. 52 (8): 489–509. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/math/0409186">math/0409186</a>. <a href="CiteSeerX_(identifier)" class="mw-redirect" title="CiteSeerX (identifier)">CiteSeerX</a> <a rel="nofollow" class="external text" href="https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.122.4429">10.1.1.122.4429</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2Ftit.2005.862083">10.1109/tit.2005.862083</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:7033413">7033413</a>.</cite>

</li>

<li id="cite_note-11"><a href="#cite_ref-11">^</a> Candès, E.J., & Wakin, M.B., An Introduction To Compressive Sampling, IEEE Signal Processing Magazine, V.21, March 2008 <a rel="nofollow" class="external autonumber" href="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4472240&isnumber=4472102">[1]</a>

</li>

<li id="cite_note-12"><a href="#cite_ref-12">^</a> Stefan Rolewicz. Metric Linear Spaces.

</li>

<li id="cite_note-14"><a href="#cite_ref-14">^</a> <a rel="nofollow" class="external text" href="http://www.acm.caltech.edu/l1magic/">L1-MAGIC is a collection of MATLAB routines</a>

</li>

<li id="cite_note-EPTV-15">^ <a href="#cite_ref-EPTV_15-0">a</a> <a href="#cite_ref-EPTV_15-1">b</a> <a href="#cite_ref-EPTV_15-2">c</a> <cite id="CITEREFTianJiaYuanPan2011" class="citation journal cs1">Tian, Z.; Jia, X.; Yuan, K.; Pan, T.; Jiang, S. B. (2011). <a rel="nofollow" class="external text" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4026331">"Low-dose CT reconstruction via edge preserving total variation regularization"</a>. Phys Med Biol. 56 (18): 5949–5967. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/1009.2288">1009.2288</a>. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2011PMB....56.5949T">2011PMB....56.5949T</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1088%2F0031-9155%2F56%2F18%2F011">10.1088/0031-9155/56/18/011</a>. <a href="PMC_(identifier)" class="mw-redirect" title="PMC (identifier)">PMC</a> <a rel="nofollow" class="external text" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4026331">4026331</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/21860076">21860076</a>.</cite>

</li>

<li id="cite_note-Orientation_and_directional_refinement-16">^ <a href="#cite_ref-Orientation_and_directional_refinement_16-0">a</a> <a href="#cite_ref-Orientation_and_directional_refinement_16-1">b</a> <a href="#cite_ref-Orientation_and_directional_refinement_16-2">c</a> <cite id="CITEREFXuan_FeiZhihui_WeiLiang_Xiao2013" class="citation journal cs1">Xuan Fei; Zhihui Wei; Liang Xiao (2013). "Iterative Directional Total Variation Refinement for Compressive Sensing Image Reconstruction". IEEE Signal Processing Letters. 20 (11): 1070–1073. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2013ISPL...20.1070F">2013ISPL...20.1070F</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2FLSP.2013.2280571">10.1109/LSP.2013.2280571</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:8156085">8156085</a>.</cite>

</li>

<li id="cite_note-Original_source_for_IRLS-17">^ <a href="#cite_ref-Original_source_for_IRLS_17-0">a</a> <a href="#cite_ref-Original_source_for_IRLS_17-1">b</a> <cite id="CITEREFCandesWakinBoyd2008" class="citation journal cs1">Candes, E. J.; Wakin, M. B.; Boyd, S. P. (2008). "Enhancing sparsity by reweighted l1 minimization". J. Fourier Anal. Appl. 14 (5–6): 877–905. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/0711.1612">0711.1612</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1007%2Fs00041-008-9045-x">10.1007/s00041-008-9045-x</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:5879257">5879257</a>.</cite>

</li>

<li id="cite_note-Iteration-18"><a href="#cite_ref-Iteration_18-0">^</a> Lange, K.: Optimization, Springer Texts in Statistics. Springer, New York (2004)

</li>

<li id="cite_note-Forward-Backward-19"><a href="#cite_ref-Forward-Backward_19-0">^</a> <cite id="CITEREFCombettesWajs2005" class="citation journal cs1">Combettes, P; Wajs, V (2005). "Signal recovery by proximal forward-backward splitting". Multiscale Model Simul. 4 (4): 1168–200. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1137%2F050626090">10.1137/050626090</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:15064954">15064954</a>.</cite>

</li>

<li id="cite_note-CGLS-20"><a href="#cite_ref-CGLS_20-0">^</a> <cite id="CITEREFHestenesStiefel1952" class="citation journal cs1">Hestenes, M; Stiefel, E (1952). <a rel="nofollow" class="external text" href="https://doi.org/10.6028%2Fjres.049.044">"Methods of conjugate gradients for solving linear systems"</a>. Journal of Research of the National Bureau of Standards. 49 (6): 409–36. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.6028%2Fjres.049.044">10.6028/jres.049.044</a>.</cite>

</li>

<li id="cite_note-Structure_tensor-21"><a href="#cite_ref-Structure_tensor_21-0">^</a> <cite id="CITEREFBroxWeickertBurgethMrázek2006" class="citation journal cs1">Brox, T.; Weickert, J.; Burgeth, B.; Mrázek, P. (2006). "Nonlinear structure tensors". Image Vis. Comput. 24 (1): 41–55. <a href="CiteSeerX_(identifier)" class="mw-redirect" title="CiteSeerX (identifier)">CiteSeerX</a> <a rel="nofollow" class="external text" href="https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.170.6085">10.1.1.170.6085</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1016%2Fj.imavis.2005.09.010">10.1016/j.imavis.2005.09.010</a>.</cite>

</li>

<li id="cite_note-TV-22"><a href="#cite_ref-TV_22-0">^</a> <cite id="CITEREFGoldlueckeStrekalovskiyCremersSiims2012" class="citation journal cs1">Goldluecke, B.; Strekalovskiy, E.; Cremers, D.; Siims, P.-T. A. I. (2012). "The natural vectorial total variation which arises from geometric measure theory". SIAM J. Imaging Sci. 5 (2): 537–563. <a href="CiteSeerX_(identifier)" class="mw-redirect" title="CiteSeerX (identifier)">CiteSeerX</a> <a rel="nofollow" class="external text" href="https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.364.3997">10.1.1.364.3997</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1137%2F110823766">10.1137/110823766</a>.</cite>

</li>

<li id="cite_note-23"><a href="#cite_ref-23">^</a> <cite id="CITEREFAndrea_MassaPaolo_RoccaGiacomo_Oliveri2015" class="citation journal cs1">Andrea Massa; Paolo Rocca; Giacomo Oliveri (2015). "Compressive Sensing in Electromagnetics – A Review". IEEE Antennas and Propagation Magazine. 57 (1): 224–238. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2015IAPM...57..224M">2015IAPM...57..224M</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2FMAP.2015.2397092">10.1109/MAP.2015.2397092</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:30196057">30196057</a>.</cite>

</li>

<li id="cite_note-Seismic_sparse_signals-24"><a href="#cite_ref-Seismic_sparse_signals_24-0">^</a> <cite id="CITEREFTaylorBanksMcCoy1979" class="citation journal cs1">Taylor, H.L.; Banks, S.C.; McCoy, J.F. (1979). "Deconvolution with the 1 norm". Geophysics. 44 (1): 39–52. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1190%2F1.1440921">10.1190/1.1440921</a>.</cite>

</li>

<li id="cite_note-LASSO-25"><a href="#cite_ref-LASSO_25-0">^</a> <cite id="CITEREFTibshirani1996" class="citation journal cs1">Tibshirani, R (1996). <a rel="nofollow" class="external text" href="http://www-stat.stanford.edu/~tibs/lasso/lasso.pdf">"Regression shrinkage and selection via the lasso"</a> (PDF). J. R. Stat. Soc. B. 58 (1): 267–288. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1111%2Fj.2517-6161.1996.tb02080.x">10.1111/j.2517-6161.1996.tb02080.x</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:16162039">16162039</a>.</cite>

</li>

<li id="cite_note-26"><a href="#cite_ref-26">^</a> <cite id="CITEREFDavid_Schneider2013" class="citation magazine cs1">David Schneider (March 2013). <a rel="nofollow" class="external text" href="https://spectrum.ieee.org/camera-chip-makes-alreadycompressed-images">"New Camera Chip Captures Only What It Needs"</a>. IEEE Spectrum. Retrieved 2013-03-20.</cite>

</li>

<li id="cite_note-cscamera-27"><a href="#cite_ref-cscamera_27-0">^</a> <cite class="citation web cs1"><a rel="nofollow" class="external text" href="https://web.archive.org/web/20100605170550/http://dsp.rice.edu/cscamera">"Compressive Imaging: A New Single-Pixel Camera"</a>. Rice DSP. Archived from <a rel="nofollow" class="external text" href="http://dsp.rice.edu/cscamera">the original</a> on 2010-06-05. Retrieved 2013-06-04.</cite>

</li>

<li id="cite_note-28"><a href="#cite_ref-28">^</a> <cite class="citation web cs1"><a rel="nofollow" class="external text" href="https://web.archive.org/web/20160120072320/http://www.technologyreview.com/view/515651/bell-labs-invents-lensless-camera/">"Bell Labs Invents Lensless Camera"</a>. MIT Technology Review. 2013-05-25. Archived from <a rel="nofollow" class="external text" href="http://www.technologyreview.com/view/515651/bell-labs-invents-lensless-camera/">the original</a> on 2016-01-20. Retrieved 2013-06-04.</cite>

</li>

<li id="cite_note-29"><a href="#cite_ref-29">^</a> <cite id="CITEREFGang_HuangHong_JiangKim_MatthewsPaul_Wilford2013" class="citation conference cs1">Gang Huang; Hong Jiang; Kim Matthews; Paul Wilford (2013). Lensless Imaging by Compressive Sensing. 2013 IEEE International Conference on Image Processing. Vol. 2393. pp. 2101–2105. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/1305.7181">1305.7181</a>. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2013arXiv1305.7181H">2013arXiv1305.7181H</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2FICIP.2013.6738433">10.1109/ICIP.2013.6738433</a>. <a href="ISBN_(identifier)" class="mw-redirect" title="ISBN (identifier)">ISBN</a> <bdi>978-1-4799-2341-0</bdi>.</cite>

</li>

<li id="cite_note-30"><a href="#cite_ref-30">^</a> <cite id="CITEREFBradyChoiMarksHorisaki2009" class="citation journal cs1">Brady, David; Choi, Kerkil; Marks, Daniel; Horisaki, Ryoichi; Lim, Sehoon (2009). <a rel="nofollow" class="external text" href="https://doi.org/10.1364%2Foe.17.013040">"Compressive holography"</a>. Optics Express. 17 (15): 13040–13049. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2009OExpr..1713040B">2009OExpr..1713040B</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1364%2Foe.17.013040">10.1364/oe.17.013040</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/19654708">19654708</a>.</cite>

</li>

<li id="cite_note-31"><a href="#cite_ref-31">^</a> <cite id="CITEREFRivensonSternJavidi2010" class="citation journal cs1">Rivenson, Y.; Stern, A.; Javidi, B. (2010). "Compressive fresnel holography". Journal of Display Technology. 6 (10): 506–509. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2010JDisT...6..506R">2010JDisT...6..506R</a>. <a href="CiteSeerX_(identifier)" class="mw-redirect" title="CiteSeerX (identifier)">CiteSeerX</a> <a rel="nofollow" class="external text" href="https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.391.2020">10.1.1.391.2020</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2Fjdt.2010.2042276">10.1109/jdt.2010.2042276</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:7460759">7460759</a>.</cite>

</li>

<li id="cite_note-32"><a href="#cite_ref-32">^</a> <cite id="CITEREFDenisLorenzThibautFournier2009" class="citation journal cs1">Denis, Loic; Lorenz, Dirk; Thibaut, Eric; Fournier, Corinne; Trede, Dennis (2009). <a rel="nofollow" class="external text" href="https://hal-ujm.archives-ouvertes.fr/ujm-00397994/file/OL_sparse_reconstruction_DH_revised_2line.pdf">"Inline hologram reconstruction with sparsity constraints"</a> (PDF). Opt. Lett. 34 (22): 3475–3477. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2009OptL...34.3475D">2009OptL...34.3475D</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1364%2Fol.34.003475">10.1364/ol.34.003475</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/19927182">19927182</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:14377881">14377881</a>.</cite>

</li>

<li id="cite_note-33"><a href="#cite_ref-33">^</a> <cite id="CITEREFMarimAngeliniOlivo-MarinAtlan2011" class="citation journal cs1">Marim, M.; Angelini, E.; Olivo-Marin, J. C.; Atlan, M. (2011). "Off-axis compressed holographic microscopy in low-light conditions". Optics Letters. 36 (1): 79–81. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/1101.1735">1101.1735</a>. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2011OptL...36...79M">2011OptL...36...79M</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1364%2Fol.36.000079">10.1364/ol.36.000079</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/21209693">21209693</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:24074045">24074045</a>.</cite>

</li>

<li id="cite_note-34"><a href="#cite_ref-34">^</a> <cite id="CITEREFMarimAtlanAngeliniOlivo-Marin2010" class="citation journal cs1">Marim, M. M.; Atlan, M.; Angelini, E.; Olivo-Marin, J. C. (2010). "Compressed sensing with off-axis frequency-shifting holography". Optics Letters. 35 (6): 871–873. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/1004.5305">1004.5305</a>. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2010OptL...35..871M">2010OptL...35..871M</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1364%2Fol.35.000871">10.1364/ol.35.000871</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/20237627">20237627</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:9738556">9738556</a>.</cite>

</li>

<li id="cite_note-35"><a href="#cite_ref-35">^</a> <cite id="CITEREFFernandez_CullWiknerMaitMattheiss2010" class="citation journal cs1">Fernandez Cull, Christy; Wikner, David A.; Mait, Joseph N.; Mattheiss, Michael; Brady, David J. (2010). "Millimeter-wave compressive holography". Appl. Opt. 49 (19): E67 – E82. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2010ApOpt..49E..67C">2010ApOpt..49E..67C</a>. <a href="CiteSeerX_(identifier)" class="mw-redirect" title="CiteSeerX (identifier)">CiteSeerX</a> <a rel="nofollow" class="external text" href="https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.1018.5231">10.1.1.1018.5231</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1364%2Fao.49.000e67">10.1364/ao.49.000e67</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/20648123">20648123</a>.</cite>

</li>

<li id="cite_note-36"><a href="#cite_ref-36">^</a> <cite class="citation magazine cs1"><a rel="nofollow" class="external text" href="https://www.wired.com/science/discoveries/news/2008/03/new_face_recognition">"Engineers Test Highly Accurate Face Recognition"</a>. <a href="Wired_(magazine)" title="Wired (magazine)">Wired</a>. 2008-03-24. <a rel="nofollow" class="external text" href="https://web.archive.org/web/20140110024041/https://www.wired.com/science/discoveries/news/2008/03/new_face_recognition">Archived</a> from the original on 2014-01-10.</cite>

</li>

<li id="cite_note-dx.doi.org-37"><a href="#cite_ref-dx.doi.org_37-0">^</a> <cite id="CITEREFLustig2007" class="citation journal cs1">Lustig, Michael (2007). <a rel="nofollow" class="external text" href="https://doi.org/10.1002%2Fmrm.21391">"Sparse MRI: The application of compressed sensing for rapid MR imaging"</a>. Magnetic Resonance in Medicine. 58 (6): 1182–1195. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1002%2Fmrm.21391">10.1002/mrm.21391</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/17969013">17969013</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:15370510">15370510</a>.</cite>

</li>

<li id="cite_note-Compressed_Sensing_MRI_2008-38"><a href="#cite_ref-Compressed_Sensing_MRI_2008_38-0">^</a> <cite id="CITEREFLustigDonohoSantosPauly2008" class="citation journal cs1">Lustig, M.; Donoho, D.L.; Santos, J.M.; Pauly, J.M. (2008). "Compressed Sensing MRI;". IEEE Signal Processing Magazine. 25 (2): 72–82. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2008ISPM...25...72L">2008ISPM...25...72L</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2FMSP.2007.914728">10.1109/MSP.2007.914728</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:945906">945906</a>.</cite>

</li>

<li id="cite_note-39"><a href="#cite_ref-39">^</a> <cite id="CITEREFEllenberg2010" class="citation magazine cs1"><a href="Jordan_Ellenberg" title="Jordan Ellenberg">Ellenberg, Jordan</a> (2010-03-04). <a rel="nofollow" class="external text" href="https://www.wired.com/2010/02/ff-algorithm/">"Fill in the Blanks: Using Math to Turn Lo-Res Datasets Into Hi-Res Samples"</a>. Wired. Vol. 18, no. 3. Retrieved 2024-04-20.</cite>

</li>

<li id="cite_note-40"><a href="#cite_ref-40">^</a> <cite id="CITEREFZhangPeterson2014" class="citation journal cs1">Zhang, Y.; Peterson, B. (2014). <a rel="nofollow" class="external text" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4058219">"Energy Preserved Sampling for Compressed Sensing MRI"</a>. Computational and Mathematical Methods in Medicine. 2014: 546814. <a href="ArXiv_(identifier)" class="mw-redirect" title="ArXiv (identifier)">arXiv</a>:<a rel="nofollow" class="external text" href="https://arxiv.org/abs/1501.03915">1501.03915</a>. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2015CMMM.201514104T">2015CMMM.201514104T</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1155%2F2014%2F546814">10.1155/2014/546814</a>. <a href="PMC_(identifier)" class="mw-redirect" title="PMC (identifier)">PMC</a> <a rel="nofollow" class="external text" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4058219">4058219</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/24971155">24971155</a>.</cite>

</li>

<li id="cite_note-Zhang_2015-41"><a href="#cite_ref-Zhang_2015_41-0">^</a> <cite id="CITEREFZhang2015" class="citation journal cs1">Zhang, Y. (2015). "Exponential Wavelet Iterative Shrinkage Thresholding Algorithm for Compressed Sensing Magnetic Resonance Imaging". Information Sciences. 322: 115–132. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1016%2Fj.ins.2015.06.017">10.1016/j.ins.2015.06.017</a>.</cite>

</li>

<li id="cite_note-42"><a href="#cite_ref-42">^</a> <cite id="CITEREFZhangWang2015" class="citation journal cs1">Zhang, Y.; Wang, S. (2015). "Exponential Wavelet Iterative Shrinkage Thresholding Algorithm with Random Shift for Compressed Sensing Magnetic Resonance Imaging". IEEJ Transactions on Electrical and Electronic Engineering. 10 (1): 116–117. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1002%2Ftee.22059">10.1002/tee.22059</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:109854375">109854375</a>.</cite>

</li>

<li id="cite_note-MRI-43"><a href="#cite_ref-MRI_43-0">^</a> <cite id="CITEREFFigueiredoBioucas-DiasNowak2007" class="citation journal cs1">Figueiredo, M.; Bioucas-Dias, J.M.; Nowak, R.D. (2007). <a rel="nofollow" class="external text" href="https://zenodo.org/record/896391">"Majorization–minimization algorithms for wavelet-based image restoration"</a>. IEEE Trans. Image Process. 16 (12): 2980–2991. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2007ITIP...16.2980F">2007ITIP...16.2980F</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1109%2Ftip.2007.909318">10.1109/tip.2007.909318</a>. <a href="PMID_(identifier)" class="mw-redirect" title="PMID (identifier)">PMID</a> <a rel="nofollow" class="external text" href="https://pubmed.ncbi.nlm.nih.gov/18092597">18092597</a>. <a href="S2CID_(identifier)" class="mw-redirect" title="S2CID (identifier)">S2CID</a> <a rel="nofollow" class="external text" href="https://api.semanticscholar.org/CorpusID:8160052">8160052</a>.</cite>

</li>

<li id="cite_note-44"><a href="#cite_ref-44">^</a> [Network tomography via compressed sensing|<a rel="nofollow" class="external free" href="http://www.ee.washington.edu/research/funlab/Publications/2010/CS-Tomo.pdf">http://www.ee.washington.edu/research/funlab/Publications/2010/CS-Tomo.pdf</a>]

</li>

<li id="cite_note-45"><a href="#cite_ref-45">^</a> <cite class="citation web cs1 cs1-prop-unfit"><a rel="nofollow" class="external text" href="https://web.archive.org/web/20130331045434/http://inviewcorp.com/products/">"InView web site"</a>. inviewcorp.com. Archived from the original on 2013-03-31.</cite>

</li>

<li id="cite_note-46"><a href="#cite_ref-46">^</a> <a rel="nofollow" class="external text" href="https://web.archive.org/web/20160916125614/http://mnras.oxfordjournals.org/content/395/3/1733">Compressed sensing imaging techniques for radio interferometry</a>

</li>

<li id="cite_note-47"><a href="#cite_ref-47">^</a> <cite id="CITEREFStevensKovarikAbellanYuan2015" class="citation journal cs1">Stevens, Andrew; Kovarik, Libor; Abellan, Patricia; Yuan, Xin; Carin, Lawrence; Browning, Nigel D. (13 August 2015). <a rel="nofollow" class="external text" href="https://doi.org/10.1186%2Fs40679-015-0009-3">"Applying compressive sensing to TEM video: a substantial frame rate increase on any camera"</a>. Advanced Structural and Chemical Imaging. 1 (1). <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1186%2Fs40679-015-0009-3">10.1186/s40679-015-0009-3</a>.</cite>

</li>

<li id="cite_note-48"><a href="#cite_ref-48">^</a> <cite id="CITEREFKovarikStevensLiyuBrowning2016" class="citation journal cs1">Kovarik, L.; Stevens, A.; Liyu, A.; Browning, N. D. (17 October 2016). <a rel="nofollow" class="external text" href="https://zenodo.org/record/1232079">"Implementing an accurate and rapid sparse sampling approach for low-dose atomic resolution STEM imaging"</a>. Applied Physics Letters. 109 (16): 164102. <a href="Bibcode_(identifier)" class="mw-redirect" title="Bibcode (identifier)">Bibcode</a>:<a rel="nofollow" class="external text" href="https://ui.adsabs.harvard.edu/abs/2016ApPhL.109p4102K">2016ApPhL.109p4102K</a>. <a href="Doi_(identifier)" class="mw-redirect" title="Doi (identifier)">doi</a>:<a rel="nofollow" class="external text" href="https://doi.org/10.1063%2F1.4965720">10.1063/1.4965720</a>.</cite>

</li>

</ol></div>

<div class="mw-heading mw-heading2"><h2 id="Further_reading">Further reading</h2></div>

<ul><li>"The Fundamentals of Compressive Sensing" <a rel="nofollow" class="external text" href="http://www.brainshark.com/brainshark/brainshark.net/portal/title.aspx?pid=zCdz10BfTRz0z0">Part 1</a>, <a rel="nofollow" class="external text" href="http://www.brainshark.com/brainshark/brainshark.net/portal/title.aspx?pid=zCgzXgcEKz0z0">Part 2</a> and <a rel="nofollow" class="external text" href="http://www.brainshark.com/brainshark/brainshark.net/portal/title.aspx?pid=zAvz9F41cz0z0">Part 3</a>: video tutorial by Mark Davenport, Georgia Tech. at <a rel="nofollow" class="external text" href="http://www.brainshark.com/sps">SigView, the IEEE Signal Processing Society Tutorial Library</a>.</li>

<li><a rel="nofollow" class="external text" href="https://www.wired.com/magazine/2010/02/ff_algorithm/all/1">Using Math to Turn Lo-Res Datasets Into Hi-Res Samples</a> Wired Magazine article</li>

<li><a rel="nofollow" class="external text" href="http://arquivo.pt/wayback/20160516193158/http://dsp.rice.edu/cs/">Compressive Sensing Resources</a> at <a href="Rice_University" title="Rice University">Rice University</a>.</li>

<li><a rel="nofollow" class="external text" href="https://www.ams.org/happening-series/hap7-pixel.pdf">Compressed Sensing Makes Every Pixel Count</a> – article in the AMS What's Happening in the Mathematical Sciences series</li>

<li><a rel="nofollow" class="external text" href="https://web.archive.org/web/20150504060355/http://ugcs.caltech.edu/~srbecker/wiki/Main_Page">Wiki on sparse reconstruction</a></li></ul>

/* start https://en.wikipedia.org/ */

.mw-parser-output .hlist dl,.mw-parser-output .hlist ol,.mw-parser-output .hlist ul{margin:0;padding:0}.mw-parser-output .hlist dd,.mw-parser-output .hlist dt,.mw-parser-output .hlist li{margin:0;display:inline}.mw-parser-output .hlist.inline,.mw-parser-output .hlist.inline dl,.mw-parser-output .hlist.inline ol,.mw-parser-output .hlist.inline ul,.mw-parser-output .hlist dl dl,.mw-parser-output .hlist dl ol,.mw-parser-output .hlist dl ul,.mw-parser-output .hlist ol dl,.mw-parser-output .hlist ol ol,.mw-parser-output .hlist ol ul,.mw-parser-output .hlist ul dl,.mw-parser-output .hlist ul ol,.mw-parser-output .hlist ul ul{display:inline}.mw-parser-output .hlist .mw-empty-li{display:none}.mw-parser-output .hlist dt::after{content:": "}.mw-parser-output .hlist dd::after,.mw-parser-output .hlist li::after{content:" · ";font-weight:bold}.mw-parser-output .hlist dd:last-child::after,.mw-parser-output .hlist dt:last-child::after,.mw-parser-output .hlist li:last-child::after{content:none}.mw-parser-output .hlist dd dd:first-child::before,.mw-parser-output .hlist dd dt:first-child::before,.mw-parser-output .hlist dd li:first-child::before,.mw-parser-output .hlist dt dd:first-child::before,.mw-parser-output .hlist dt dt:first-child::before,.mw-parser-output .hlist dt li:first-child::before,.mw-parser-output .hlist li dd:first-child::before,.mw-parser-output .hlist li dt:first-child::before,.mw-parser-output .hlist li li:first-child::before{content:" (";font-weight:normal}.mw-parser-output .hlist dd dd:last-child::after,.mw-parser-output .hlist dd dt:last-child::after,.mw-parser-output .hlist dd li:last-child::after,.mw-parser-output .hlist dt dd:last-child::after,.mw-parser-output .hlist dt dt:last-child::after,.mw-parser-output .hlist dt li:last-child::after,.mw-parser-output .hlist li dd:last-child::after,.mw-parser-output .hlist li dt:last-child::after,.mw-parser-output .hlist li li:last-child::after{content:")";font-weight:normal}.mw-parser-output .hlist ol{counter-reset:listitem}.mw-parser-output .hlist ol>li{counter-increment:listitem}.mw-parser-output .hlist ol>li::before{content:" "counter(listitem)"\a0 "}.mw-parser-output .hlist dd ol>li:first-child::before,.mw-parser-output .hlist dt ol>li:first-child::before,.mw-parser-output .hlist li ol>li:first-child::before{content:" ("counter(listitem)"\a0 "}

/* end https://en.wikipedia.org/ */

</style><style data-mw-deduplicate="TemplateStyles:r1236075235">

/* start https://en.wikipedia.org/ */

.mw-parser-output .navbox{box-sizing:border-box;border:1px solid #a2a9b1;width:100%;clear:both;font-size:88%;text-align:center;padding:1px;margin:1em auto 0}.mw-parser-output .navbox .navbox{margin-top:0}.mw-parser-output .navbox+.navbox,.mw-parser-output .navbox+.navbox-styles+.navbox{margin-top:-1px}.mw-parser-output .navbox-inner,.mw-parser-output .navbox-subgroup{width:100%}.mw-parser-output .navbox-group,.mw-parser-output .navbox-title,.mw-parser-output .navbox-abovebelow{padding:0.25em 1em;line-height:1.5em;text-align:center}.mw-parser-output .navbox-group{white-space:nowrap;text-align:right}.mw-parser-output .navbox,.mw-parser-output .navbox-subgroup{background-color:#fdfdfd}.mw-parser-output .navbox-list{line-height:1.5em;border-color:#fdfdfd}.mw-parser-output .navbox-list-with-group{text-align:left;border-left-width:2px;border-left-style:solid}.mw-parser-output tr+tr>.navbox-abovebelow,.mw-parser-output tr+tr>.navbox-group,.mw-parser-output tr+tr>.navbox-image,.mw-parser-output tr+tr>.navbox-list{border-top:2px solid #fdfdfd}.mw-parser-output .navbox-title{background-color:#ccf}.mw-parser-output .navbox-abovebelow,.mw-parser-output .navbox-group,.mw-parser-output .navbox-subgroup .navbox-title{background-color:#ddf}.mw-parser-output .navbox-subgroup .navbox-group,.mw-parser-output .navbox-subgroup .navbox-abovebelow{background-color:#e6e6ff}.mw-parser-output .navbox-even{background-color:#f7f7f7}.mw-parser-output .navbox-odd{background-color:transparent}.mw-parser-output .navbox .hlist td dl,.mw-parser-output .navbox .hlist td ol,.mw-parser-output .navbox .hlist td ul,.mw-parser-output .navbox td.hlist dl,.mw-parser-output .navbox td.hlist ol,.mw-parser-output .navbox td.hlist ul{padding:0.125em 0}.mw-parser-output .navbox .navbar{display:block;font-size:100%}.mw-parser-output .navbox-title .navbar{float:left;text-align:left;margin-right:0.5em}body.skin--responsive .mw-parser-output .navbox-image img{max-width:none!important}@media print{body.ns-0 .mw-parser-output .navbox{display:none!important}}

/* end https://en.wikipedia.org/ */

</style></div><div role="navigation" class="navbox authority-control" aria-labelledby="Authority_control_databases_frameless&#124;text-top&#124;10px&#124;alt=Edit_this_at_Wikidata&#124;link=https&#58;//www.wikidata.org/wiki/Q2665459#identifiers&#124;class=noprint&#124;Edit_this_at_Wikidata723" style="padding:3px"><table class="nowraplinks hlist mw-collapsible autocollapse navbox-inner" style="border-spacing:0;background:transparent;color:inherit"><tbody><tr><th scope="col" class="navbox-title" colspan="2"><div id="Authority_control_databases_frameless&#124;text-top&#124;10px&#124;alt=Edit_this_at_Wikidata&#124;link=https&#58;//www.wikidata.org/wiki/Q2665459#identifiers&#124;class=noprint&#124;Edit_this_at_Wikidata723" style="font-size:114%;margin:0 4em">Authority control databases </div></th></tr><tr><th scope="row" class="navbox-group" style="width:1%">National</th><td class="navbox-list-with-group navbox-list navbox-odd" style="width:100%;padding:0"><div style="padding:0 0.25em"><ul><li><a rel="nofollow" class="external text" href="https://d-nb.info/gnd/1036377245">Germany</a></li><li><a rel="nofollow" class="external text" href="https://id.loc.gov/authorities/sh2015001812">United States</a></li><li><a rel="nofollow" class="external text" href="https://www.nli.org.il/en/authorities/987007414138105171">Israel</a></li></ul></div></td></tr><tr><th scope="row" class="navbox-group" style="width:1%">Other</th><td class="navbox-list-with-group navbox-list navbox-even" style="width:100%;padding:0"><div style="padding:0 0.25em"><ul><li><a rel="nofollow" class="external text" href="https://lux.collections.yale.edu/view/concept/1d42abbb-4fbd-4fe9-b9dc-d7391078f8ec">Yale LUX</a></li></ul></div></td></tr></tbody></table></div></div><div><div class="zim-footer">

This article is issued from <a class="external text" title="Last edited on 2025-08-03" href="https://en.wikipedia.org/wiki/?title=Compressed_sensing&oldid=1303991080">Wikipedia</a>. The text is available under <a class="external text" href="https://creativecommons.org/licenses/by-sa/4.0/deed.en">Creative Commons Attribution-Share Alike 4.0</a> unless otherwise noted. Additional terms may apply for the media files.

</div>

</div></div>

</div>

</main>

</div>

</body></html>