Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

Source function from two-particle correlation function through entropy-regularized Richardson-Lucy deblurring

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A maximum-entropy regularized Richardson-Lucy algorithm recovers the source function from two-particle correlations while strictly preserving its normalization.

desk verdict A useful normalization-preserving deblurring method for femtoscopy source extraction, with a real but fixable selection-bias issue in hyperparameter tuning. read the letter →

arxiv 2502.09478 v1 pith:AKWOOD24 submitted 2025-02-13 nucl-th

C. K. Tam , Z. Chajęcki , P. Danielewicz , P. Nzabahimana This is my paper · ORCID
classification nucl-th
keywords Richardson-LucydeblurringKoonin-Prattequationsourcefunctionreconstructionmaximumentropyregularizationtwo-particlecorrelationsheavy-ioncollisionsproton-protondeuteron-alpha
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

Two-particle correlation functions $C(q)$ encode the size and shape of the particle-emitting region in heavy-ion collisions, but recovering the source function $S(r)$ from them means inverting the Koonin-Pratt equation, an ill-posed deconvolution. This paper proposes MEM-RL, a Richardson-Lucy deblurring scheme augmented with Lucy's maximum-entropy regularization, and demonstrates that it restores $S(r)$ while strictly conserving the integral of the source, which total-variation regularization does not. The method is tested on a known Gaussian source and on experimental proton-proton and deuteron-alpha correlations, returning source sizes and purities consistent with Gaussian fits and Bayesian imaging. The practical payoff is a shape-unbiased route from measured correlations to emission-source geometry.

What carries the argument

The load-bearing object is the MEM-RL update rule $\Delta S_j = \nu(\Delta H_j + \Delta S_j^{\mathrm{ent}})$: $\Delta H_j$ is the standard Richardson-Lucy correction $\Delta H_j = S_j(\sum_i \tilde{C}_i K_{ij}/C_i - 1)$, and $\Delta S_j^{\mathrm{ent}} = -\alpha S_j(S_j + \ln(S_j/\chi_j) + 1 - \sum_k (S_k/\chi_k)\Pi_{kj})$ is Lucy's maximum-entropy term, with the floating default $\chi_j = \sum_k \Pi_{jk} S_k$ built from a Gaussian-smearing kernel. Because each term separately satisfies $\sum_j \Delta H_j = 0$ and $\sum_j \Delta S_j^{\mathrm{ent}} = 0$, the total update conserves $\sum_j S_j$; combined with the acceleration coefficient $\nu = 1.99$ and a positivity-preserving initial guess, this keeps the iteration stable and normalized. The regularization strength $\alpha$, smearing width $\sigma_r$, and purity $\lambda$ are optimized by $\chi^2$ minimization, with detector resolution folded in as an extra Gaussian smearing $\sigma_q$ for the deuteron-$\alpha$ kernel.

What would settle it

Generate a correlation function from a known source that violates the smoothness assumption, e.g., a momentum-dependent $S(q,r)$ representing an expanding collective source, apply MEM-RL to the angle-averaged $C(q)$, and compare the recovered $S(r)$ with the true angle-averaged source; a significant mismatch in size or purity would show that the method cannot be trusted under collective motion. Alternatively, simulate a sharp-edged or double-peaked $S(r)$ with known normalization and noise; if MEM-RL cannot reproduce the structure without introducing unphysical $r^2S(r)$ oscillations, the claim that the regularization prevents overfitting while preserving normalization fails.

Watch

Extended reading notes

Core claim

The central claim is that the entropy-regularized update combines the standard Richardson-Lucy data-driven step with a maximum-entropy smoothing step in a way that always sums to zero over the source, so normalization is preserved exactly at every iteration. The algorithm therefore avoids the divergence that can occur with total-variation regularization when the smoothing parameter is too large. On a Gaussian test case with $R_G=2.5$ fm and $\lambda=0.65$, MEM-RL recovers an FWHM of $4.25\pm0.61$ fm on blurred data and a purity of $0.64\pm0.04$; for experimental $pp$ correlations it gives $r_{1/2}=4.17^{+0.38}_{-0.38}$ fm and $\lambda=0.66^{+0.05}_{-0.05}$, matching the Gaussian fit and Bayesian imaging results, and for $d$--$\alpha$ correlations it gives $r_{1/2}=5.39\pm0.52$ fm with $\lambda=0.98\pm0.02$, again matching the Gaussian fit. These agreements are taken as evidence that MEM-RL recovers both the shape and the normalization of the underlying source.

Load-bearing premise

The angle-averaged Koonin-Pratt equation assumes the source function is the same at every pair momentum; if emission depends on momentum, the recovered source function is not the quantity the correlation function encodes.

Editorial extensions

If this is right

  • A user of MEM-RL does not need to assume a Gaussian or any other parametric source shape; the algorithm returns a full $S(r)$ profile, so non-Gaussian tails and structures can be seen directly.
  • Because normalization is conserved by construction, the extracted purity parameter $\lambda$ can be read off the recovered source rather than enforced as a separate constraint.
  • The two-hyperparameter optimization ($\sigma_r$ and $\lambda$, with $\alpha$ fixed) is enough in practice, since $\chi^2$ is roughly independent of $\alpha$ once the other two are optimized.
  • When the emission source is momentum-dependent, e.g., under collective motion, the Koonin-Pratt equation becomes a Hadamard product and this deblurring framework does not apply; the recovered $S(r)$ would then be misleading.
  • The scheme extends to non-identical pairs with sharp resonance peaks, such as deuteron-alpha, provided the kernel includes detector resolution via a Gaussian smearing.

Reading between the lines

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

  • One could apply the same entropy-regularized update to other positive linear inverse problems in nuclear physics, such as deprojecting emission profiles or unfolding invariant-mass spectra, where positivity and normalization are both physical requirements.
  • Since the method is non-parametric, it is well suited to searching for halo-like tails or secondary-decay contributions that Gaussian fits would smooth over; the paper's own tail comparison with Bayesian imaging hints at this capability but does not pursue it.
  • A direct stress test would be to feed MEM-RL correlations generated from a sharply structured source, such as two well-separated radii, and check whether the recovered $S(r)$ splits cleanly; the paper only validates against smooth Gaussian sources.
  • The reported insensitivity to $\alpha$ may be specific to the noise level and binning used here; testing across a range of statistical errors would show whether two-parameter optimization remains safe at low statistics.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes a maximum-entropy-regularized Richardson-Lucy algorithm (MEM-RL) for inverting the angle-averaged Koonin-Pratt equation, i.e. for restoring the two-particle source function S(r) from a measured correlation function C(q). The algorithm combines the standard RL multiplicative update with Lucy's entropy regularization and a floating-default prior constructed by Gaussian smearing of the current estimate; the authors emphasize that this combination preserves the normalization of S(r), unlike earlier total-variation-regularized RL. The method is tested on a simulated Gaussian C_pp, on experimental C_pp from 40Ca+40Ca at 80 MeV/nucleon, and on a preliminary d-alpha correlation function from the HiRA10 experiment. The extracted FWHM source sizes and purity parameters are compared with Gaussian fits and, for C_pp, with Bayesian imaging, and are reported in Table I and Figs. 4-6.

Significance. If the method is accepted as stated, it provides a useful addition to the source-function extraction toolkit: it is non-parametric, preserves source normalization by construction, and is benchmarked against both a known Gaussian truth in simulation and against established methods (Gaussian fitting and Bayesian imaging) on real data. The d-alpha application with explicit detector-resolution smearing also demonstrates the method on a kernel with sharp resonances. The central result is not circular: for the simulated case the algorithm recovers the true source size within uncertainties, and for real data the results are consistent with independent methods. The main caveat is that the hyperparameter optimization includes a post-hoc rejection of solutions deemed 'unphysical', and the reported bootstrap uncertainties are conditioned on that filter; this weakens the claim that the quoted errors reflect the full inference procedure. The paper would be significantly strengthened by quantifying the sensitivity to the filter and by making the acceptance criterion and all normalization conditions fully explicit.

major comments (3)
  1. [Section III B, Fig. 3] The optimization procedure explicitly discards restored sources that show multiple peaks in r^2 S(r) or non-vanishing tails, and this filter is applied not only to the Gaussian simulation but also to the experimental data. Since the filter is aligned with the known truth in the simulation but is a subjective regularity assumption for the data, the reported r_1/2 values in Table I and Figs. 4-5 are conditional on this acceptance rule. In particular, because the filter removes a region of large-lambda solutions, the inferred lambda and source size are partly determined by the filter rather than by the data alone. Please quantify this sensitivity: e.g. report results with and without the filter, or include the filter in the bootstrap procedure, and justify the criterion a priori rather than after inspecting the solutions.
  2. [Section III B, bootstrap paragraph] The bootstrap uncertainty estimate as described samples the correlation function from N(C(q), delta C(q)) and then applies MEM-RL 'with an optimized parameter set', but it is not stated whether the hyperparameters (lambda, alpha, sigma_r) are re-optimized for each bootstrap sample and whether the unphysical-solution filter is applied to each sample. As written, the quoted 1-sigma and 2-sigma bands exclude the selection step and therefore understate the total uncertainty. Please specify the bootstrap protocol precisely and, if the filter is applied per sample, document how many samples are rejected and how that affects the quoted intervals.
  3. [Section II C, Eq. (14)] The claim that Eq. (14) strictly preserves source normalization, i.e. that the sum of Delta S_j^S vanishes, is asserted in the text but the normalization conditions on the smearing matrix Pi are not stated precisely. The result holds if Pi is symmetric and doubly stochastic (so that sum_j S_j Pi_kj = chi_k), but Eq. (13) only says 'properly normalized' and 'usually symmetric'. Since normalization preservation is the central advantage claimed over TV regularization, please state the exact condition on Pi and give the short derivation of sum_j Delta S_j^S = 0.
minor comments (5)
  1. [Eq. (7)] The symbol K is used both for the continuum kernel K(q,r) and for the discretized matrix K_ij; this is confusing in Eqs. (6)-(8), especially because K_ij contains a 1/lambda term. Please use distinct notation, e.g. a boldface or calligraphic symbol for the matrix.
  2. [Eq. (9)] The normalization formulas appear corrupted: 'Sj → Sj/Sj = S (norm)/ X j Sj' is not readable. Please rewrite the definitions of S_j^{(norm)} and the scaling of S by C^{(norm)}/K^{(norm)} in a clean equation block.
  3. [Section III A] The paper states N=25 and M=13 for q in (0,100) MeV/c, but the pp kernel in Fig. 1 and the data applications use q in (0,80) MeV/c with 2 MeV/c bins (40 bins). Please clarify which convention is used in the standard-RL test and in Fig. 4, and ensure the dimensional constraint N >= M is stated consistently.
  4. [Section II D, Eq. (15)] The acceleration parameter nu = 1.99 is introduced without explaining why this value is 'generally a good choice' or how it was selected. Please state the range of nu for which the update is guaranteed to keep the source non-negative, or cite a reference for the convergence properties.
  5. [Section IV, d-alpha discussion] The sentence introducing the detector-resolution smearing says the rest of the procedure is identical to the proton-proton case, but the forward model is changed by sigma_q and the kernel for d-alpha is energy-dependent and complex. Please spell out how the kernel smearing is implemented in Eq. (7)-(8), especially how the normalization of the smeared kernel is preserved.

Circularity Check

0 steps flagged · score 0.0 of 10

No load-bearing circularity: MEM-RL is validated against a known simulated source and cross-checked with independent Gaussian-fit and Bayesian-imaging estimates; the self-citations are background, not the derivation.

full rationale

The paper's central derivation chain is self-contained. The MEM-RL update (Eq. 15) combines the standard RL step (Eq. 10) with Lucy's entropy regularization (Eq. 14), and the claimed normalization preservation is a direct algebraic property of those equations, not a fitted result. The method is tested on a simulated Gaussian source with known parameters, and the recovered size r1/2 = 4.25 ± 0.61 fm is compared with the true value 4.16 fm, so the validation uses external truth rather than the paper's own inputs. For experimental data, the extracted r1/2 and λ are compared with independent Gaussian chi-square fits and Bayesian imaging from Ref. [4]; this agreement is a consistency check, not a consequence of the model definition. Several citations are to the authors' prior work, but none is invoked as a uniqueness theorem or as a substitute for the derivation; the d–alpha kernel is ultimately tied to measured phase shifts [37]. The post-hoc rejection of restored sources with multiple peaks or non-vanishing tails (Sec. III B) is a selection rule that can narrow bootstrap uncertainties, but it does not by itself fix the peak position or width, so the recovered source sizes still carry information from the data. The paper also honestly flags the absence of a definitive hyperparameter metric and the difficulty of accurate small-r restoration, which are limitations rather than circular steps.

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

The inversion relies on the KP forward model with an exactly known kernel, three hand-tuned hyperparameters (lambda, alpha, sigma_r), and an added detector-resolution parameter for d-alpha. No new physical entities are introduced. The method is not circular because it is benchmarked against known truth in simulation and against independent imaging and Gaussian estimates on real data.

free parameters (5)
  • purity lambda = 0.64 +/- 0.04 (simulation); 0.66 +0.05/-0.05 (pp data); 0.98 +/- 0.02 (d-alpha data)
    Fraction of correlated pairs, absorbed into the kernel normalization in Eq. (7) and optimized via chi-square grid search. It controls the source amplitude and tail.
  • regularization strength alpha = 0.3 fixed for d-alpha; otherwise searched on a grid in (0,1)
    Weights the entropy penalty against the data-driven term in Eq. (15). The paper notes there is no definitive metric for choosing it.
  • floating-default smoothing width sigma_r = 1.07 +0.24/-0.24 fm (pp); 1.27 +/- 0.13 fm (d-alpha)
    Width of the Gaussian smearing matrix Pi in Eq. (13) that defines the prior. It controls smoothness and is chosen by chi-square minimization.
  • detector resolution smearing sigma_q = 2.2 MeV/c (d-alpha)
    Gaussian width added to the kernel to model the experimental resolution that broadens the narrow 6Li resonance peak. Fitted as an extra parameter for the d-alpha case.
  • r-space binning configuration = N=25, M=13 with r in (0,39) fm; uneven binning for d-alpha
    Manually chosen to satisfy the dimensionality constraint (N >= M) and to avoid divergence at r=0. The choice affects the restored source shape, especially for kernels with sharp features.
assumptions (5)
  • domain assumption The angle-averaged Koonin-Pratt equation (Eq. 1) relates C(q) to a source S(r) independent of q under the smoothness assumption.
    Invoked in the Introduction; this is the forward model for the entire inversion. The authors later state (Section IV) that collective motion, where S depends on q, is outside the framework.
  • domain assumption The two-particle kernel K(q,r) is known exactly from solving the Schrodinger equation with phase-shift-matched potentials.
    The deconvolution is non-blind; the kernel is treated as exact. No uncertainty in the kernel is propagated into the bootstrap errors, and in the simulation test the same kernel is used for generation and restoration.
  • standard math The Richardson-Lucy updates preserve non-negativity and the total integral when the kernel is properly normalized.
    A standard property of the RL algorithm, used to justify its suitability for positive counting data.
  • ad hoc to paper The entropy regularization term from Lucy (1994), Eq. (14), preserves the source normalization.
    Taken from the literature; the paper asserts that P_j DeltaS_Sj = 0 but does not prove it in the text.
  • ad hoc to paper The floating default prior, a Gaussian-smoothed version of the current source estimate, is a valid approximation of the unknown source.
    This is a modeling choice standard in entropy-based image restoration. The width sigma_r is fitted, but the Gaussian functional form is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Source function from two-particle correlation function through entropy-regularized Richardson-Lucy deblurring." pith.science (2026). https://pith.science/paper/AKWOOD24

@misc{pith2026250209478,
  author       = {Pith},
  title        = {Pith review of: Source function from two-particle correlation function through entropy-regularized Richardson-Lucy deblurring},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AKWOOD24}},
  note         = {Machine review of arXiv:2502.09478}
}
abstract

Source functions are obtained from $p$-$p$ and $d$--$\alpha$ correlation functions by applying the Richardson-Lucy (RL) deblurring to the Koonin-Pratt (KP) equation. To prevent fitting of noise in the correlation function, total-variation (TV) regularization is employed that has been effective in ordinary image restoration. TV alone cannot ensure normalization of the source functions. To ensure the latter, we propose a maximum-entropy regularized RL algorithm (MEM-RL). We outline the MEM-RL formalism and optimization strategy for the KP equation, demonstrating its effectiveness on both simulated and experimental data, including the $p$-$p$ and $d$-$\alpha$ correlation functions.

Figures

Figures reproduced from arXiv: 2502.09478 by the authors.

Figure 1
Figure 1. FIG. 1. Kernels employed in the Koonin-Pratt equation plotted in a logarithmic scale. The upper [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Restoration of the Gaussian source function from its associated smooth correlation function [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Optimization test on the parameter set ( [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Results of MEM-RL applied to correlation function calculated from a Gaussian source [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Results of MEM-RL applied to experimental correlation function [4]. Panels (a) and (b) [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Results of the MEM-RL algorithm (blue and green bands) and [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Femtoscopy Measurement with S$\pi$RIT TPC in Radioactive BeamHeavy-ion Collisions

    physics.ins-det 2026-07 conditional novelty 5.0 of 10

    A first proton-proton femtoscopy correlation function measured with the SπRIT TPC in 132Sn+124Sn collisions at 270 MeV/u shows the expected attractive-interaction peak near 20 MeV/c.

  2. Transport Theory and Correlation Measurements: Coming to Terms on Emission Sources

    nucl-th 2025-06 conditional novelty 4.0 of 10

    A two-parameter exponential tail added to BUU-computed proton sources reproduces measured p-p correlations, but the tail parameters are fit to the same data they describe.

Reference graph

Works this paper leans on

38 extracted references · 34 canonical work pages · cited by 2 Pith papers

  1. [1]

    D. Q. Fang, Y. G. Ma, X. Y. Sun, P. Zhou, Y. Togano, N. Aoi, H. Baba, X. Z. Cai, X. G. Cao, J.G.Chen, Y.Fu, W.Guo, Y.Hara, T.Honda, Z.G.Hu, K.Ieki, Y.Ishibashi, Y.Ito, N.Iwasa, S. Kanno, T. Kawabata, H. Kimura, Y. Kondo, K. Kurita, M. Kurokawa, T. Moriguchi, H. Murakami, H. Ooishi, K. Okada, S. Ota, A. Ozawa, H. Sakurai, S. Shimoura, R. Shioda, E. Takeshi...

    work page 2016

  2. [2]

    Verde, D

    G. Verde, D. A. Brown, P. Danielewicz, C. K. Gelbke, W. G. Lynch, and M. B. Tsang, Phys. Rev. C 65, 054609 (2002)

    work page 2002

  3. [3]

    Verde, P

    G. Verde, P. Danielewicz, W. Lynch, C. Chan, C. Gelbke, L. Kwong, T. Liu, X. Liu, D. Sey- mour, R. Shomin, W. Tan, M. Tsang, A. Wagner, H. Xu, D. Brown, B. Davin, Y. Larochelle, R. de Souza, R. Yanez, R. Charity, and L. Sobotka, Physics Letters B653, 12 (2007)

    work page 2007

  4. [4]

    Henzl, M

    V. Henzl, M. A. Kilburn, Z. Chajęcki, D. Henzlova, W. G. Lynch, D. Brown, A. Chbihi, D. D. S. Coupland, P. Danielewicz, R. T. deSouza, M. Famiano, C. Herlitzius, S. Hudan, J. Lee, S. Lukyanov, A. M. Rogers, A. Sanetullaev, L. G. Sobotka, Z. Y. Sun, M. B. Tsang, A. Vander Molen, G. Verde, M. S. Wallace, and M. Youngs, Phys. Rev. C85, 014606 (2012)

    work page 2012

  5. [5]

    Y. Wang, F. Guan, Q. Wu, X. Diao, Y. Huang, L. Lyu, Y. Qin, Z. Qin, D. Si, Z. Bai, F. Duan, L. Duan, Z. Gao, Q. Hu, R. Hu, G. Jin, S. Jin, J. Ma, P. Ma, J. Wang, P. Wang, Y. Wang, X. Wei, H. Yang, Y. Yang, G. Yu, Y. Yu, Y. Zhang, Q. Zhou, Y. Zhang, C. Ma, X. Hu, H. Wang, Y. Cui, J. Tian, and Z. Xiao, Physics Letters B825, 136856 (2022). 19

    work page 2022

  6. [6]

    F. Zhu, W. G. Lynch, T. Murakami, C. K. Gelbke, Y. D. Kim, T. K. Nayak, R. Pelak, M. B. Tsang, H. M. Xu, W. G. Gong, W. Bauer, K. Kwiatkowski, R. Płaneta, S. Rose, V. E. Viola, L. W. Woo, S. Yennello, and J. Zhang, Phys. Rev. C44, R582 (1991)

    work page 1991

  7. [7]

    Pratt, Phys

    S. Pratt, Phys. Rev. C56, 1095 (1997)

    work page 1997

  8. [8]

    S. E. Koonin, Physics Letters B70, 43 (1977)

    1977

Show all 38 references
  1. [9]

    Pratt, T

    S. Pratt, T. Csörgő, and J. Zimányi, Phys. Rev. C42, 2646 (1990)

    1990

  2. [10]

    Pratt and M

    S. Pratt and M. B. Tsang, Phys. Rev. C36, 2390 (1987)

    1987

  3. [11]

    W. G. Gong, W. Bauer, C. K. Gelbke, and S. Pratt, Phys. Rev. C43, 781 (1991)

    1991

  4. [12]

    Messiah,Quantum Mechanics, Quantum Mechanics No

    A. Messiah,Quantum Mechanics, Quantum Mechanics No. v.2 (Elsevier Science, 1961)

    1961

  5. [13]

    Kopylov, Physics Letters B50, 472 (1974)

    G. Kopylov, Physics Letters B50, 472 (1974)

    1974

  6. [14]

    M. A. Lisa, W. G. Gong, C. K. Gelbke, and W. G. Lynch, Phys. Rev. C44, 2865 (1991)

    1991

  7. [15]

    D. A. Brown and P. Danielewicz, Phys. Rev. C64, 014902 (2001)

    2001

  8. [16]

    D. A. Brown, A. Enokizono, M. Heffner, R. Soltz, P. Danielewicz, and S. Pratt, Phys. Rev. C 72, 054902 (2005)

    2005

  9. [17]

    S. Y. Panitkin, N. N. Ajitanand, J. Alexander, M. Anderson, D. Best, F. P. Brady, T. Case, W. Caskey, D. Cebra, J. Chance, P. Chung, B. Cole, K. Crowe, A. Das, J. Draper, M. Gilkes, S. Gushue, M. Heffner, A. Hirsch, E. Hjort, L. Huo, M. Justice, M. Kaplan, D. Keane, J. Kint- n...

    2001

  10. [18]

    Chung, N

    P. Chung, N. N. Ajitanand, J. M. Alexander, M. Anderson, D. Best, F. P. Brady, T. Case, W. Caskey, D. Cebra, J. L. Chance, B. Cole, K. Crowe, A. C. Das, J. E. Draper, M. L. Gilkes, S. Gushue, M. Heffner, A. S. Hirsch, E. L. Hjort, W. Holzmann, L. Huo, M. Issah, M. Justice, M. ...

    2003

  11. [19]

    W. H. Richardson, J. Opt. Soc. Am.62, 55 (1972)

    1972

  12. [20]

    L. B. Lucy, Astronomical Journal79, 745 (1974)

    1974

  13. [21]

    Y.Sakai, S.Yamada, T.Sato, R.Hayakawa, R.Higurashi,andN.Kominato,TheAstrophysical Journal 951, 59 (2023)

    2023

  14. [22]

    W. B. Atwood, A. A. Abdo, M. Ackermann, W. Althouse, B. Anderson, M. Axelsson, L. Bal- dini, J. Ballet, D. L. Band, G. Barbiellini, J. Bartelt, D. Bastieri, B. M. Baughman, K. Bechtol, D. Bédérède, F. Bellardi, R. Bellazzini, B. Berenji, G. F. Bignami, D. Bisello, E. Bissaldi,...

    2009

  15. [23]

    Prato, M., Cavicchioli, R., Zanni, L., Boccacci, P., and Bertero, M., A&A539, A133 (2012)

    2012

  16. [24]

    L., Meyer, S., Sarli, E., and Bartelmann, M., A&A553, A118 (2013)

    Konrad, S., Majer, C. L., Meyer, S., Sarli, E., and Bartelmann, M., A&A553, A118 (2013)

    2013

  17. [25]

    Nzabahimana, T

    P. Nzabahimana, T. Redpath, T. Baumann, P. Danielewicz, P. Giuliani, and P. Guèye, Phys. Rev. C 107, 064315 (2023)

    2023

  18. [26]

    Danielewicz and M

    P. Danielewicz and M. Kurata-Nishimura, Phys. Rev. C105, 034608 (2022)

    2022

  19. [27]

    Nzabahimana and P

    P. Nzabahimana and P. Danielewicz, Physics Letters B846, 138247 (2023)

    2023

  20. [28]

    L. I. Rudin, S. Osher, and E. Fatemi, Physica D: Nonlinear Phenomena60, 259 (1992)

    1992

  21. [29]

    L. B. Lucy, Astronomy and Astrophysics289, 983 (1994)

    1994

  22. [30]

    Sweany, W

    S. Sweany, W. Lynch, K. Brown, A. Anthony, Z. Chajecki, D. Dell’Aquila, P. Morfouace, F. Teh, C. Tsang, M. Tsang, R. Wang, and K. Zhu, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 1018, 165798 (2021)

    2021

  23. [31]

    150 (1985)

    Interacting Binaries, NATO Advanced Study Institute (ASI) Series C, Vol. 150 (1985)

    1985

  24. [32]

    Danielewicz and S

    P. Danielewicz and S. Pratt, Phys. Rev. C75, 034907 (2007)

    2007

  25. [33]

    Pratt and S

    S. Pratt and S. Petriconi, Phys. Rev. C68, 054901 (2003)

    2003

  26. [34]

    Chajęcki and M

    Z. Chajęcki and M. Lisa, Phys. Rev. C78, 064903 (2008)

    2008

  27. [35]

    Nzabahimana, P

    P. Nzabahimana, P. Danielewicz, and G. Verde, IL NUOVO CIMENTO100, 48 (2025)

    2025

  28. [36]

    Nzabahimana, Particle Correlations in Heavy-Ion Collisions (Michigan State University, 2023)

    P. Nzabahimana, Particle Correlations in Heavy-Ion Collisions (Michigan State University, 2023)

    2023

  29. [37]

    D. H. Boal, C. K. Gelbke, and B. K. Jennings, Rev. Mod. Phys.62, 553 (1990)

    1990

  30. [38]

    Chajęcki and M

    Z. Chajęcki and M. Lisa, Phys. Rev. C79, 034908 (2009). 22

    2009

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.