How to make CLEAN variants faster? Using clustered components informed by the autocorrelation function

Hendrik M\"uller; Sanjay Bhatnagar

REVIEW 3 major objections 4 minor 60 references

How to make CLEAN variants faster? Using clustered components informed by the autocorrelation function

T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Autocorr-CLEAN aims to deconvolve radio interferometer images up to an order of magnitude faster than classical CLEAN by using the residual's autocorrelation to fit clusters of components in each iteration.

desk verdict A genuinely new autocorrelation-driven CLEAN variant with a plausible large speed-up; the claim needs measured k/l ratios and a fix to a duplicated timing figure before it is established. read the letter →

arxiv 2504.16058 v1 pith:J7BEM7XZ submitted 2025-04-22 astro-ph.IM astro-ph.GA

Hendrik Müller , Sanjay Bhatnagar This is my paper · ORCID

classification astro-ph.IMastro-ph.GA

keywords radiointerferometrydeconvolutionCLEANmultiscaleimagingautocorrelationfunctionadaptivebasisfunctionscomputationalcomplexityaperturesynthesis

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Autocorr-CLEAN is a new multiscale variant of the standard CLEAN deconvolution algorithm for radio interferometry. The paper claims it reaches the reconstruction quality and iteration-level convergence of modern multiscale CLEAN methods, such as Asp-CLEAN, while keeping every inner-loop operation as cheap as classical CLEAN: shifts, rescales, and subtractions only, with no convolutions or Fourier evaluations inside the loops. The mechanism is to fit a basis function from the autocorrelation function of the current residual, then subtract many point-like components at once. On synthetic aperture-synthesis observations of three bright radio sources, the authors report reaching the same residual level as CLEAN in five to ten times less wall-clock time, and continuing to recover fainter diffuse emission. This matters because deconvolution is the innermost loop of radio data reduction; a faster minor loop translates directly into handling the data volumes of upcoming arrays.

What carries the argument

The load-bearing object is the autocorrelation function of the residual, $\Pi = I_D \diamond I_D$, treated as a second deconvolution problem that can be solved quickly with CLEAN. From its cleaned components the algorithm builds a basis function $\omega = \sum_i (\delta^{\omega_i})^\gamma$, a cloud of point components that encodes the global spatial correlation structure of the emission and is point-symmetric by construction; this cloud is the multiscale "component" subtracted in the exterior minor loop. Keeping the loops cheap requires a set of precomputed correlation products, including $M_B = \omega \star B_D$ and $M_I = \omega \star I_D$, which are updated by second-order formulas such as $\Pi \leftarrow \Pi - I_D \diamond (M \star \delta_k) - (M \star \delta_k) \diamond I_D + M \diamond M$ whenever a component is subtracted. The complexity analysis summarizes the cost as $8kN + (m/k)(5N + 8Nl) \approx mN \cdot 8l/k$, where $m$ is the number of classical CLEAN iterations, $k$ the number of components in the basis, and $l$ the number of subminor-loop iterations; the claimed speed-up rests on the assumption $l \ll k \ll m$.

What would settle it

A concrete test is to run Autocorr-CLEAN on a source with low dynamic range between bright compact emission and faint extended emission and measure $l$ (subminor-loop iterations), $k$ (components in the basis), and $m$ (iterations needed by classical CLEAN) with $N$ the number of pixels: if $l$ is not much smaller than $k$, or $k$ not much smaller than $m$, the predicted cost $mN(8l/k)$ exceeds CLEAN's $Nm$ and the wall-clock advantage disappears.

Watch

Extended reading notes

Core claim

The central claim, stated on the paper's own terms, is that a multiscale CLEAN can be made both adaptive and fast if the shape of the model component is learned from the autocorrelation function of the residual rather than from a fixed dictionary or a local optimization. The authors write the dirty image as $I_D = B_D \star I$ and its autocorrelation as $\Pi = I_D \diamond I_D$, which obeys $\Pi = (B_D \diamond B_D) \star (I \diamond I)$; so the autocorrelation problem can itself be solved with CLEAN. A subminor loop CLEANs this autocorrelation with the point-spread-function autocorrelation and collects the resulting delta components into a basis function $\omega = \sum_i (\delta^{\omega_i})^\gamma$, with $\gamma = 2$ as a Gaussian-motivated default. Each exterior minor-loop iteration then performs one MS-CLEAN step with this continuously changing, possibly non-radially symmetric basis function, subtracts many components at once, and updates all autocorrelation products analytically to second order, avoiding explicit convolutions. The paper reports that the number of iterations needed matches the adaptive-scale Asp-CLEAN algorithm, while every iteration stays at linear complexity; on the synthetic image-plane benchmarks the residual reaches CLEAN's level in five to ten times less time, and the reconstruction quality matches or improves on the comparison multiscale methods for very diffuse emission.

Load-bearing premise

The speed advantage collapses unless refreshing the autocorrelation model takes far fewer inner iterations than the number of components subtracted at once, and far fewer components than the thousands of single-component steps classical CLEAN would need.

Editorial extensions

If this is right

If the speed-up holds on real data, the deconvolution minor loop will no longer dominate pipeline runtime: the same residual level is reached five to ten times faster than with classical CLEAN, and fainter diffuse emission is recovered.
Iteration counts match Asp-CLEAN without its per-iteration nonlinear optimization, so modern multiscale reconstruction quality becomes available at near-classical per-iteration cost.
Because the basis function is re-derived from the residual autocorrelation at every step, it adapts continuously and can be non-radially symmetric, which lets the algorithm represent flat, elliptical diffuse structures that Gaussian-based multiscale CLEAN misses.
The two-loop structure fits the standard major-loop/minor-loop architecture, so the method can be dropped into existing pipelines with modest changes; the authors report implementing it in the same code library used for the comparisons.
The recommended aggressive control parameters for the gain and the autocorrelation-fitting fraction are a stated risk in poorly calibrated or low-dynamic-range data, where the authors note the underlying assumptions may be violated.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The recursive second-order update of the autocorrelation products is essentially a low-rank update of a residual-covariance estimate; the same update trick could accelerate other matching-pursuit or sparse-imaging methods that subtract localized components while keeping a global correlation model.
The authors treat $\gamma$ as a fixed scale bias; an extension they do not explore is to make $\gamma$ adaptive to the measured convergence rate per scale, which could reduce sensitivity to control-parameter choices at low dynamic range.
Because the benchmarks are image-plane only, a natural next test is end-to-end runs that include gridding, self-calibration, and flagging; the relative speed-up may shrink if the minor loop is no longer the dominant cost, or grow if fewer major loops are needed.
The idea is not tied to radio interferometry: any deconvolution problem with a known point-spread function and a residual whose autocorrelation changes slowly could use the same component-clustering heuristic.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a real new algorithm, not a repackaging. Autocorr-CLEAN uses the residual autocorrelation to build a cluster of CLEAN components as a continuously adapting, possibly non-radial basis function, updates it in a subminor loop, and keeps every inner step at linear complexity. That specific combination—autocorrelation-informed clustering plus linear-complexity updates—is new relative to Offringa & Smirnov, Asp-CLEAN, and DoB-CLEAN. The paper is candid that this is proof-of-concept: image-plane only, no interaction with gridding, calibration, or flagging, and no claim of super-resolution.

What it does well: the pseudocode in Table 1 is specific enough to reimplement; the complexity bookkeeping in Sec. 3.4 is transparent; the synthetic benchmarks cover several source morphologies and compare against a purpose-built C++ CLEAN baseline in the same library, which is more careful than most algorithm papers. The convergence plots show Autocorr-CLEAN reaching CLEAN-level residuals in roughly a fifth to a tenth of the wall time, and it handles faint diffuse emission well in some examples. The citation pattern is appropriate: the earlier subminor-loop, adaptive-scale, and non-radial-basis work is credited.

Where it is soft: (1) The complexity argument in Sec. 3.4 predicts a speed-up of order k/(8l), but no measured values of k, l, or m are reported. The paper argues l < k qualitatively; that alone gives no quantitative bound, and the text concedes the advantage can collapse at low dynamic range. The measured wall-clock speed-up is evidence that the ratio is favorable in these examples, but the theory is not yet tied to the data. (2) The headline benchmarks are single runs with no error bars, and gamma, f, and the gain were tuned at least partly on the same synthetic sets used for the headline numbers. Acceptable for a proof-of-concept, but not for a magnitude claim. (3) Fig. A.5 appears to have a copy-paste error: Asp-CLEAN at gain 0.5 shows exactly the same cumulative timings as Autocorr-CLEAN at gain 0.5. That needs fixing. (4) The CLEAN baseline is a custom reimplementation, not production CASA; the fairness case is plausible, but a comparison against an optimized reference would strengthen it.

Net: this deserves a serious referee. I would send it to review and ask for measured k, l, and m per source, repeated runs or run-to-run variability, and a corrected Fig. A.5. If those land, it is a useful pipeline-relevant contribution for the ngVLA/SKA era.

Referee Report

3 major / 4 minor

Summary. The manuscript introduces Autocorr-CLEAN, a multiscale deconvolution algorithm for radio interferometry. The algorithm fits a cloud of CLEAN components to the autocorrelation function of the residual in a subminor loop and uses the powered cloud as a basis function for a CLEAN minor-loop step, thereby avoiding explicit convolutions. The authors claim that all inner-loop operations have linear complexity, that the convergence speed in terms of iterations matches Asp-CLEAN, and that the algorithm is up to an order of magnitude faster than classical Högbom CLEAN. The method is implemented in the LibRA/CASA environment and tested on synthetic VLA observations of Cygnus A, Hercules A, and M106.

Significance. The core idea is attractive and timely: deriving a continuously adapting, potentially non-radially symmetric basis from the residual autocorrelation while keeping every inner-loop operation at linear complexity is a genuine contribution to the CLEAN family. If the speed and fidelity claims hold, the algorithm could be integrated into existing CASA-like pipelines with modest effort. The complexity bookkeeping in Sec. 3.4 is transparent, and the choice of CLEAN and Asp-CLEAN as baselines is well motivated. However, the central speed-up claim is currently not supported by measured complexity parameters, and one figure contains duplicated timing data, so the empirical case needs strengthening.

major comments (3)

[§3.4, §4.3, §5] The headline speed-up is not connected to the complexity analysis. The cost estimate in §3.4 gives Autocorr-CLEAN ≈ mN(8l/k) against CLEAN's Nm, so the predicted speed-up is k/(8l). A ten-fold speed-up therefore requires k/l ≳ 80, but the paper only argues that "fewer iterations are performed in the subminor loop than the number of components in the basis function ω" (§4.3), i.e. l < k. No measured values of l, k, or m are reported for any of the three benchmark sources. Because §3.1 and §5 concede that the advantage may collapse at low dynamic range, this omission is load-bearing: without l/k data the abstract's "up to a magnitude faster" claim is an extrapolation rather than a demonstrated result. Please report l, k, and m for each test case and verify the k/l ratio, or revise the claimed speed-up accordingly.
[Fig. A.5] The timing rows for Asp-CLEAN at gain 0.5 are identical to those for Autocorr-CLEAN at gain 0.5 (21.0, 38.0, 54.0, 74.0, 87.0, 93.0, 99.0, 104.0, 110.0 s). This is implausible because the two algorithms have very different per-iteration costs and indicates a copy-paste error. This affects the gain-robustness comparison discussed in §4.4. Please regenerate the Asp-CLEAN timing data for gain 0.5 or remove that row from the figure.
[§4.3–4.4] The empirical support rests on single-run convergence curves without error bars, and the control parameters γ, f, and gain are evaluated on the same synthetic data sets used for the headline benchmarks. Since the central claim is a quantitative speed-up, please provide either multiple noise realisations or an explicit statement that the reported curves are single realisations and that the final parameter choices (e.g., γ = 2, f = 0.1, gain = 0.3) were selected on these same data. Without this information the reader cannot assess how much of the reported speed-up is due to parameter tuning.

minor comments (4)

[Fig. 2 caption] The caption lists "Cygnus A, Hydra A, Hercules A and M106", but the text in §4.1 only describes Cygnus A, Hercules A, and M106; please clarify whether Hydra A is a fourth test case or a typo.
[Table 1, Block 6] The update equation for M̃MBB_{j+1} appears to have a self-referential right-hand side; it should presumably read M̃MBB_{j+1} = M̃MBB_j + ... . Please check and correct.
[§3.1] The text states "theoretically several orders of magnitude of speed-up over CLEAN", while the Abstract and §5 claim "up to a magnitude"; please make these statements consistent.
[§3.3] The name "Högböm" is a typo for "Högbom".

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the speed and fidelity claims are benchmarked against external baselines (Högbom CLEAN and Asp-CLEAN), and the adaptive autocorrelation-based basis function is an iterative feedback scheme rather than a prediction forced by construction.

full rationale

The derivation chain of Autocorr-CLEAN is self-contained in the relevant sense. The algorithm defines its basis function from the autocorrelation of the current residual, but this is an explicit iterative update rule, not a hidden identification of the output with an input: the basis function is recomputed from the residual and then used to update that same residual in a standard minor-loop/subminor-loop structure. No parameter is fitted to a subset of the benchmark data and then reported as a prediction of that subset; the convergence speed and reconstruction quality are measured against external, independently implemented algorithms (Högbom CLEAN and Asp-CLEAN) on synthetic observations built from real VLA images. The complexity estimate in Sec. 3.4, 8kN + (m/k)(5N + 8Nl) ≈ mN(8l/k), is an explicit analytic bound whose validity depends on the stated assumptions l << k << m and on the absence of accumulating residual errors; the paper itself flags that these assumptions may fail at low dynamic range and that the advantage may collapse. That is a correctness and measurement gap, not circularity: the speed-up formula is not equivalent to the benchmarked speed-up by construction, and no measured l, k, or m values are disguised as predictions. Self-citations to earlier work (Bhatnagar & Cornwell 2004; Müller & Lobanov 2023b; Hsieh & Bhatnagar 2021) are used as prior art, algorithmic inspiration, and performance comparison targets, not as an imported uniqueness theorem or as the sole justification for the central claim. The Gaussian-derived choice gamma = 2 is presented as a heuristic regularization knob with sensitivity tests, not as an ansatz whose adoption is forced by a self-citation. Therefore, under the stated hard rules, there is no exhibited reduction of the paper's results to their own inputs.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard image-plane interferometry modeling (Eq 1, Eq 3) plus a set of heuristic choices (autocorrelation-based basis selection, gamma power law, stopping fractions, and the assumed scaling l << k << m) that are justified by intuition and parameter sweeps rather than derived from first principles.

free parameters (4)

Power parameter gamma = 2 (recommended; fastest in sweep gamma=1.5-2)
Controls the distribution of component weights in the basis function; motivated by Gaussian self-convolution but acknowledged as an ad-hoc scale bias (Sec 3.3, Fig 6).
Subminor-loop fraction f = 0.1 (recommended; tested 0.1-0.3)
Stopping threshold for the autocorrelation subminor loop (Sec 3.3, Eq 7); smaller f gives deeper autocorrelation fitting at higher cost (Fig 7).
CLEAN gain = 0.3 (aggressive values fastest for Autocorr-CLEAN; 0.5 degrades the model)
Standard CLEAN gain; the paper finds robustness up to 0.3 for Autocorr-CLEAN, while CLEAN diverges at high gain (Sec 4.4, Fig A.5).
Switch-to-Hogbom threshold = unspecified ('user-defined number of iterations')
Number of consecutive small-scale triggers after which the algorithm permanently switches to classical CLEAN (Sec 3.3, Eq 6).

assumptions (5)

domain assumption The gridded dirty image is exactly a shift-invariant convolution of the true sky with the dirty beam: ID = BD * I (Eq 1), with no direction-dependent or wide-field effects.
Invoked throughout Sec 2.1 and used to derive the autocorrelation relation Eq 3; the paper explicitly restricts itself to an image-plane problem in Sec 4.1.
standard math The autocorrelation of the dirty image satisfies II = (BD diamond BD) * (I diamond I) (Eq 3), so the residual autocorrelation can be deconvolved with the same CLEAN machinery using the beam autocorrelation as PSF.
Follows from Eq 1 and the definition of the mirror-convolution operator diamond; this is the formal grounding of the basis-function estimator.
ad hoc to paper A CLEAN model of the residual autocorrelation, raised component-wise to power gamma, yields a good sparsifying basis function omega for the residual itself.
The paper calls this an 'ad-hoc approximation' (Sec 3.1) and admits the interpretation is heuristic; the entire speed and quality claim depends on this encoding of spatial correlation structure.
standard math The Gaussian self-convolution identity G_2sigma diamond G_2sigma proportional to G_sigma motivates fixing gamma=2 as a scale-bias.
Sec 3.3 uses the Gaussian result to recommend gamma=2, while acknowledging the exact value requires practical experience; the sweep in Fig 6 shows 1.5-2 perform best.
ad hoc to paper The iteration counts satisfy l << k << m and residual errors do not accumulate over exterior iterations.
Sec 3.4's complexity estimate 8kN + m/k(5N + 8Nl) improves on CLEAN's Nm only under these empirical scalings; the paper itself flags low-dynamic-range cases where this may fail.

how reviews work

0 comments

Cite this review

Pith. "Pith review of How to make CLEAN variants faster? Using clustered components informed by the autocorrelation function." pith.science (2026). https://pith.science/paper/J7BEM7XZ

@misc{pith2026250416058,
  author       = {Pith},
  title        = {Pith review of: How to make CLEAN variants faster? Using clustered components informed by the autocorrelation function},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J7BEM7XZ}},
  note         = {Machine review of arXiv:2504.16058}
}

read the original abstract

Deconvolution, imaging and calibration of data from radio interferometers is a challenging computational (inverse) problem. The upcoming generation of radio telescopes poses significant challenges to existing, and well proven data reduction pipelines due to the large data sizes expected from these experiments, and the high resolution and dynamic range. In this manuscript, we deal with the deconvolution problem. A variety of multiscalar variants to the classical CLEAN algorithm (the de-facto standard) have been proposed in the past, often outperforming CLEAN at the cost of significantly increasing numerical resources. In this work, we aim to combine some of these ideas for a new algorithm, Autocorr-CLEAN, to accelerate the deconvolution and prepare the data reduction pipelines for the data sizes expected by the upcoming generation of instruments. To this end, we propose to use a cluster of CLEAN components fitted to the autocorrelation function of the residual in a subminor loop, to derive continuously changing, and potentially non-radially symmetric, basis functions for CLEANing the residual. Autocorr-CLEAN allows for the superior reconstruction fidelity achieved by modern multiscalar approaches, and their superior convergence speed. It achieves this without utilizing any substep of super-linear complexity in the minor loops, keeping the single minor loop and subminor loop iterations at an execution time comparable to CLEAN. Combining these advantages, Autocorr-CLEAN is found to be up to a magnitude faster than the classical CLEAN procedure. Autocorr-CLEAN fits well in the algorithmic framework common for radio interferometry, making it relatively straightforward to include in future data reduction pipelines. With its accelerated convergence speed, and smaller residual, Autocorr-CLEAN may be an important asset for the data analysis in the future.

Figures

Figures reproduced from arXiv: 2504.16058 by the authors.

**Figure 2.** Gaussian convolved model images of Cygnus A, Hydra A, Hercules A and M106 (upper row), and respective reconstructions with the VLA in A configuration with CLEAN (second row), Asp-CLEAN (third row) and Autocorr-CLEAN (fourth row). at comparing the convergence speed of Autocorr-CLEAN to classical CLEAN approaches. To this end, it is beneficial to know the ground truth image for comparison. While aforementioned works… view at source ↗

**Figure 3.** Residual images for the reconstructions shown in 2. strengths afterwards with a minimization approach. In this sense, it presents the ‘best that we can do’ in the classical MS-CLEAN framework, naturally outperforming plain MS-CLEAN due to this scale adaptation step. The ground truth models, and the respective reconstructions are shown in [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: The recovered model and the respective residual for different deconvolution techniques as a function of number [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]

**Figure 5.** Figure 5: Speed of convergence for the different techniques as a function of number of iterations (top row), and computational time (bottom row). Left column: Cygnus A, middle column: Hercules A, right column: M106. the quantities MI, BI, ˜ MBI, ... ˜ in the subminor loop and …

**Figure 6.** Figure 6: Convergence curves for Autocorr-CLEAN with different values of γ for the Cygnus A example. 5. Conclusion In this manuscript, we presented a novel multiscalar CLEAN variant, named Autocorr-CLEAN. Autocorr-CLEAN draws strong inspiration from MS-CLEAN algorithms that wer…

**Figure 7.** Figure 7: Convergence curves for Autocorr-CLEAN with different values of f for the Cygnus A example. Bhatnagar, S., Rau, U., & Golap, K. 2013, ApJ, 770, 91 Carrillo, R. E., McEwen, J. D., & Wiaux, Y. 2012, MNRAS, 426, 1223 CASA Team, Bean, B., Bhatnagar, S., et al. 2022, PASP, 1…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 48 canonical work pages

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archiveprefix author booktitle chapter edition editor howpublished institution eprint journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 ...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in " " * FUNCTION format....
[3]

S., Dabbech , A., & Wiaux , Y

Aghabiglou , A., Chu , C. S., Dabbech , A., & Wiaux , Y. 2024, , 273, 3

work page 2024
[4]

L., Perley , R

Arras , P., Bester , H. L., Perley , R. A., et al. 2021, , 646, A84

work page 2021
[5]

& Cornwell , T

Bhatnagar , S. & Cornwell , T. J. 2004, , 426, 747

work page 2004
[6]

& Cornwell , T

Bhatnagar , S. & Cornwell , T. J. 2017, , 154, 197

work page 2017
[7]

J., Golap , K., & Uson , J

Bhatnagar , S., Cornwell , T. J., Golap , K., & Uson , J. M. 2008, , 487, 419

work page 2008
[8]

2021, Size-of-Computing Estimates for ngVLA Synthesis Imaging , ngVLA memo series

Bhatnagar, S., Hirart, R., & Pokorny, M. 2021, Size-of-Computing Estimates for ngVLA Synthesis Imaging , ngVLA memo series

work page 2021

Show all 60 references

[9]

2022, Baseline HPG runtime performance for imaging , ngVLA memo series

Bhatnagar, S., Madsen, F., & Robnett, J. 2022, Baseline HPG runtime performance for imaging , ngVLA memo series

2022
[10]

2013, , 770, 91

Bhatnagar , S., Rau , U., & Golap , K. 2013, , 770, 91

2013
[11]

E., McEwen , J

Carrillo , R. E., McEwen , J. D., & Wiaux , Y. 2012, , 426, 1223

2012
[12]

2022, , 134, 114501

CASA Team , Bean , B., Bhatnagar , S., et al. 2022, , 134, 114501

2022
[13]

Clark , B. G. 1980, , 89, 377

1980
[14]

Cornwell , T. J. 2008, IEEE Journal of Selected Topics in Signal Processing, 2, 793

2008
[15]

S., & Wiaux , Y

Dabbech , A., Aghabiglou , A., Chu , C. S., & Wiaux , Y. 2024, , 966, L34

2024
[16]

A., Smirnov , O

Dabbech , A., Repetti , A., Perley , R. A., Smirnov , O. M., & Wiaux , Y. 2021, , 506, 4855

2021
[17]

2024, , 964, L25

Event Horizon Telescope Collaboration , Akiyama , K., Alberdi , A., et al. 2024, , 964, L25

2024
[18]

F., Leong , O., Sun , H., & Bouman , K

Gao , A. F., Leong , O., Sun , H., & Bouman , K. L. 2023, arXiv e-prints, arXiv:2303.12217

2023 arXiv
[19]

H \"o gbom , J. A. 1974, , 15, 417

1974
[20]

C., Roth , J

Homan , D. C., Roth , J. S., & Pushkarev , A. B. 2024, , 167, 11

2024
[21]

2022 a , Recommended Developments Necessary for Applying (W)Asp Deconvolution Algorithms to ngVLA, aRDG memo series

Hsieh, G., Bhatnagar, S.and Hiriart, R., & Pokorny, M. 2022 a , Recommended Developments Necessary for Applying (W)Asp Deconvolution Algorithms to ngVLA, aRDG memo series

2022
[22]

& Bhatnagar, S

Hsieh, G. & Bhatnagar, S. 2021, Efficient Adaptive-Scale CLEAN Deconvolution in CASA for Radio Interferometric Images, aRDG memo series

2021
[23]

2022 b , An Adaptive-Scale Multi-Frequency Deconvolution of Interferometric Images, aRDG memo series

Hsieh, G., Rau, U., & Bhatnagar. 2022 b , An Adaptive-Scale Multi-Frequency Deconvolution of Interferometric Images, aRDG memo series

2022
[24]

2024, arXiv e-prints, arXiv:2406.01342

Jarret , A., Kashani , S., Ru \'e -Queralt , J., et al. 2024, arXiv e-prints, arXiv:2406.01342

2024 arXiv
[25]

R., Selig , M., & En lin , T

Junklewitz , H., Bell , M. R., Selig , M., & En lin , T. A. 2016, , 586, A76

2016
[26]

2023, arXiv e-prints, arXiv:2306.06007

Kashani , S., Ru \'e Queralt , J., Jarret , A., & Simeoni , M. 2023, arXiv e-prints, arXiv:2306.06007

2023 arXiv
[27]

2023, Imaging Unmitigated ALMA Cubes, nAASC memo series

Kepley, A., Madsen, F., Robnet, J., & Rowe, K. 2023, Imaging Unmitigated ALMA Cubes, nAASC memo series

2023
[28]

A., Tsutsumi , T., Brogan , C

Kepley , A. A., Tsutsumi , T., Brogan , C. L., et al. 2020, , 132, 024505

2020
[29]

S., et al

Kim , J.-S., Mueller , H., Nikonov , A. S., et al. 2024, arXiv e-prints, arXiv:2409.00540

2024 arXiv
[30]

u ller , H., R \

Kramer , J. A., M \"u ller , H., R \"o der , J., & Ros , E. 2025, arXiv e-prints, arXiv:2503.04970

2025 arXiv
[31]

1997, , 123, 183

Lannes , A., Anterrieu , E., & Marechal , P. 1997, , 123, 183

1997
[32]

P., Waters , B., Schiebel , D., Young , W., & Golap , K

McMullin , J. P., Waters , B., Schiebel , D., Young , W., & Golap , K. 2007, in Astronomical Society of the Pacific Conference Series, Vol. 376, Astronomical Data Analysis Software and Systems XVI, ed. R. A. Shaw , F. Hill , & D. J. Bell , 127

2007
[33]

2024, , 689, A299

M \"u ller , H. 2024, , 689, A299

2024
[34]

& Lobanov , A

M \"u ller , H. & Lobanov , A. P. 2022, , 666, A137

2022
[35]

& Lobanov , A

M \"u ller , H. & Lobanov , A. P. 2023 a , , 673, A151

2023
[36]

& Lobanov , A

M \"u ller , H. & Lobanov , A. P. 2023 b , , 672, A26

2023
[37]

2024, , 684, A47

M \"u ller , H., Massa , P., Mus , A., Kim , J.-S., & Perracchione , E. 2024, , 684, A47

2024
[38]

2023, , 675, A60

M \"u ller , H., Mus , A., & Lobanov , A. 2023, , 675, A60

2023
[39]

2024 a , , 688, A100

Mus , A., M \"u ller , H., & Lobanov , A. 2024 a , , 688, A100

2024
[40]

2024 b , , 684, A55

Mus , A., M \"u ller , H., Mart \' -Vidal , I., & Lobanov , A. 2024 b , , 684, A55

2024
[41]

R., McKinley , B., Hurley-Walker , N., et al

Offringa , A. R., McKinley , B., Hurley-Walker , N., et al. 2014, , 444, 606

2014
[42]

Offringa , A. R. & Smirnov , O. 2017, , 471, 301

2017
[43]

E., Repetti , A., et al

Onose , A., Carrillo , R. E., Repetti , A., et al. 2016, , 462, 4314

2016
[44]

2017, , 469, 938

Onose , A., Dabbech , A., & Wiaux , Y. 2017, , 469, 938

2017
[45]

F., Debbrecht , L

Paraschos , G. F., Debbrecht , L. C., Kramer , J. A., et al. 2024, , 686, L5

2024
[46]

& Cornwell , T

Rau , U. & Cornwell , T. J. 2011, , 532, A71

2011
[47]

2023, Galaxies, 11, 12

Roelofs , F., Blackburn , L., Lindahl , G., et al. 2023, Galaxies, 11, 12

2023
[48]

2023, , 678, A177

Roth , J., Arras , P., Reinecke , M., et al. 2023, , 678, A177

2023
[49]

2022, , 664, A134

Schmidt , K., Geyer , F., Fr \"o se , S., et al. 2022, , 664, A134

2022
[50]

Schwab , F. R. 1984, , 89, 1076

1984
[51]

1994, , 283, 349

Starck , J.-L., Bijaoui , A., Lopez , B., & Perrier , C. 1994, , 283, 349

1994
[52]

J., & Donoho , D

Starck , J.-L., Candes , E. J., & Donoho , D. L. 2002, IEEE Transactions on Image Processing, 11, 670

2002
[53]

L., Elad , M., & Donoho , D

Starck , J. L., Elad , M., & Donoho , D. L. 2005, IEEE Transactions on Image Processing, 14, 1570

2005
[54]

Starck , J. L. & Murtagh , F. 2006, Astronomical image and data analysis (Springer)

2006
[55]

2015, Sparse image and signal processing: Wavelets and related geometric multiscale analysis, second edition, 1--423

Starck, J.-L., Murtagh, F., & Fadili, J. 2015, Sparse image and signal processing: Wavelets and related geometric multiscale analysis, second edition, 1--423

2015
[56]

2013, , 553, A105

Tasse , C., van der Tol , S., van Zwieten , J., van Diepen , G., & Bhatnagar , S. 2013, , 553, A105

2013
[57]

R., Moran , J

Thompson , A. R., Moran , J. M., & Swenson , George W., J. 2017, Interferometry and Synthesis in Radio Astronomy, 3rd Edition

2017
[58]

Wakker , B. P. & Schwarz , U. J. 1988, , 200, 312

1988
[59]

Wiaux , Y., Jacques , L., Puy , G., Scaife , A. M. M., & Vandergheynst , P. 2009, , 395, 1733

2009
[60]

2016, , 592, A128

Zhang , L., Bhatnagar , S., Rau , U., & Zhang , M. 2016, , 592, A128

2016

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archiveprefix author booktitle chapter edition editor howpublished institution eprint journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 ...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in " " * FUNCTION format....

[3] [3]

S., Dabbech , A., & Wiaux , Y

Aghabiglou , A., Chu , C. S., Dabbech , A., & Wiaux , Y. 2024, , 273, 3

work page 2024

[4] [4]

L., Perley , R

Arras , P., Bester , H. L., Perley , R. A., et al. 2021, , 646, A84

work page 2021

[5] [5]

& Cornwell , T

Bhatnagar , S. & Cornwell , T. J. 2004, , 426, 747

work page 2004

[6] [6]

& Cornwell , T

Bhatnagar , S. & Cornwell , T. J. 2017, , 154, 197

work page 2017

[7] [7]

J., Golap , K., & Uson , J

Bhatnagar , S., Cornwell , T. J., Golap , K., & Uson , J. M. 2008, , 487, 419

work page 2008

[8] [8]

2021, Size-of-Computing Estimates for ngVLA Synthesis Imaging , ngVLA memo series

Bhatnagar, S., Hirart, R., & Pokorny, M. 2021, Size-of-Computing Estimates for ngVLA Synthesis Imaging , ngVLA memo series

work page 2021