Pith. sign in

REVIEW 3 major objections 4 minor 3 cited by

For intensity-only imaging, the optimal optical response sends each source to a single distinct detector — generalized focusing — for both Shannon-capacity and Fisher-information objectives.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-02 11:09 UTC pith:MRUUVRNN

load-bearing objection The core optimality theorem is sound and the contribution is real; the partially coherent extension is asserted on a supplement we can't see. the 3 major comments →

arxiv 2606.16724 v2 pith:MRUUVRNN submitted 2026-06-15 physics.optics

End-to-end meta-imagers: Information-theoretic objectives and generalized focusing optima

Lukas Kienesberger , Zeyu Kuang , Yaxi Liu , Owen D. Miller This is my paper
classification physics.optics
keywords computational imaginginformation theorypermutation matrixFisher informationShannon capacityintensity bottleneckmetasurface inverse designpoint spread function
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper tries to establish a fundamental limit on what the optics in a computational imager can do when detectors measure only intensity. It develops two closed-form, data-free objectives — one from Shannon capacity, one from Fisher information — that score a photonic stage without any training data. It then proves that both objectives, and any objective in a broad convex family, are maximized by a permutation matrix: each source's light must land on exactly one distinct detector. The authors call this generalized focusing, and show numerically that inverse-designed scatterers converge to it in geometries where conventional focusing intuition fails, including two-way imagers, random scattering media, and Hermite–Gauss mode sorting. They also show the same constraint holds for coherent and partially coherent sources, because the bottleneck is the nonnegativity of intensity measurements, not source coherence.

Core claim

The paper's central claim is that for intensity-only detection, the optimal incoherent transfer matrix T — for both the Shannon-capacity objective and the Fisher-information/CRB objective, and for every objective of the form Σ f(λ_j²) with f smooth, strictly convex and decreasing — is a permutation matrix. Equivalently, each source's emission is concentrated on a single, distinct detector, a condition the authors call generalized focusing. The proof relies on the observation that passive optics constrain the columns of T to be nonnegative with column sums at most 1, which places the eigenvalues of TᵀT on a simplex; strict convexity then forces all eigenvalues to 1, so TᵀT = I, and only permu

What carries the argument

The central object is the incoherent transfer matrix T mapping source intensities to detector intensities, with nonnegative entries and column sums ≤ 1. The argument runs through the singular values of T: both proposed objectives (regularized Shannon capacity C = ½ log det(I + (P²/2πeσ²)TTᵀ) and Fisher/CRB objective E = σ² Tr[(TTᵀ + SNR⁻²I)⁻¹]) are functions of the eigenvalues of TᵀT. The column-sum constraint implies ∥μ∥₁ ≤ N, so the eigenvalues lie in a simplex; strict convexity and monotonicity of f push the optimum to μ = (1,…,1), i.e., TᵀT = I. A nonnegative orthonormal matrix must be a permutation matrix. The 'intensity bottleneck' — the fact that nonnegative vectors cannot form a comp

Load-bearing premise

The proof assumes a passive optical system — every column of the transfer matrix is nonnegative and sums to at most 1 — and that the imaging is done by unbiased estimation with no prior information about the scene; if any of these fails, permutation optimality need not hold.

What would settle it

Run a numerical search over all N×N matrices with nonnegative entries and column sums ≤ 1 for N ≥ 3, evaluating the regularized Shannon capacity (Eq. 5) and the Fisher objective (Eq. 9); if any matrix beats the best permutation matrix for the same SNR and noise level, the theorem is false. Alternatively, fabricate or simulate a metasurface whose intensity transfer matrix is demonstrably non-permutation and show its reconstruction error falls below the best permutation-matrix design under identical noise and unbiased reconstruction.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • Data-free objectives can replace full end-to-end optimization: optimizing Shannon capacity or Fisher information alone yields reconstruction performance matching end-to-end training, verified through a random scattering medium.
  • Any speckled or mixing point-spread function carries a scaling penalty: if a single input produces M hot spots, output resolution scales as N/M rather than N.
  • The generalized-focusing optimum holds regardless of source/detector geometry, including non-planar arrangements such as the two-way imager.
  • The constraint persists for coherent and partially coherent inputs with intensity-only detection: only permutation-like unitary matrices (up to output/input phases) allow phase-independent full amplitude recovery.
  • With unequal source/detector counts, generalized focusing persists (with unused detectors for S<N, or multiple sources sharing each detector for S>N), so only N independent channels can be resolved without priors.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The theorem suggests a concrete design rule for computational metasurface imagers: unless prior information is exploited, the photonic stage should be engineered to be as close to permutation-like as possible; attempts to design informative spread-out point-spread functions for intensity cameras are fighting a fundamental limit.
  • If the intensity bottleneck is the true root, interferometric or phase-preserving detection at the focal plane should be able to circumvent generalized focusing and unlock mixing designs for coherent sources — a testable prediction that extends the paper's discussion.
  • The convexity argument also implies that the optimal transfer matrix is independent of the specific noise level (as long as it is Gaussian and the regularization is set by SNR), so designs should transfer across noise conditions; this is a corollary worth verifying in practice.
  • The same nonnegativity obstruction suggests that any intensity-only measurement scheme claiming to recover more degrees of freedom than the permutation optimum must be introducing priors or bias, which the paper's framework would classify as prior-aware extensions rather than violations of the limit.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes two deterministic, data-free objectives for designing the photonic stage of an intensity-only computational imager: a Shannon-capacity surrogate (Eq. 5) and a Fisher-information/CRB surrogate (Eq. 9), both closed-form functions of the incoherent transfer matrix T. It proves that under nonnegativity and column-sum constraints (passive optics), the optima of both objectives—and a broader family of singular-value cost functions—are permutation matrices, i.e., each source is focused onto a single distinct detector ('generalized focusing'). This is validated in three inverse-design settings (two-way imager, random-scatterer imaging, Hermite-Gauss mode sorting), with the Fisher objective matching end-to-end optimization. The paper further claims that this focusing constraint is independent of source coherence, arguing via a coherent-field amplitude-recovery proof and a Fisher-information Schur-complement argument deferred to the SM.

Significance. If the claims hold, the paper provides a strong, counterintuitive design principle: for unbiased, prior-free intensity imaging, the information-optimal photonic response is non-mixing and delta-like, regardless of source/detector geometry. The incoherent proof is elegant and essentially correct: the column-sum constraint confines the eigenvalues of T^T T to a simplex, convexity forces equality, and nonnegative orthonormal columns are necessarily permutation vectors. The data-free objectives are practical, and the numerical comparison to end-to-end optimization is a genuine strength. However, the advertised coherence-independence result is not established in the submitted manuscript, and a stated general-family theorem has a sign/direction error; these must be fixed before the broadest claims can be accepted.

major comments (3)
  1. [§V, after Eq. (14)] The abstract and Sec. IV rely on the claim that generalized focusing persists for partially coherent sources, but the only support in the main text is the sentence 'A complementary Fisher-information analysis confirms the same conclusion (cf. SM)' with no SM included. This is load-bearing: the coherent proof preceding it treats only fully coherent fields through a unitary T, and the partially coherent intensity model is linear in the coherence-matrix elements. If some coherence or phase parameters are treated as unknown nuisances, the Schur-complement argument is not immediate; non-permutation T could in principle allow amplitude recovery after marginalization. Please include the proof in the main text or SM, or clearly restrict the claim to fully coherent sources.
  2. [§III, 'One can widen the class...'] The theorem for the wider family is stated as 'every function of the singular values of the form F = Σ_j f(λ_j²) with f smooth, strictly convex, and monotonically decreasing is optimized by a permutation matrix.' As written, this is false for maximization: if f is decreasing, maximizing Σ f(µ_j) over the simplex pushes all µ_j to 0, not to 1. The proof sketch actually establishes that such F is minimized at µ_j=1. The Shannon capacity objective enters as the negative of such an F. Please restate the theorem as a minimization claim and clarify the sign convention for the Shannon case.
  3. [§III proof; §VI discussion] The proof optimizes over the full set of nonnegative matrices with column sums ≤1 and concludes that 'any permutation' is optimal. Physical passive reciprocal systems may not realize arbitrary permutations: when source and detector ports coincide, reciprocity forces T symmetric, and a symmetric permutation matrix is an involution. The paper's examples use distinct input/output ports, so they are consistent, but the universal 'regardless of geometry' claim needs a caveat or a proof that arbitrary permutations are feasible in the considered geometries. Otherwise the mathematical optimum and the physically realizable optimum may differ.
minor comments (4)
  1. [§V] The symbol T is used both for the incoherent intensity transfer matrix (Eq. (1)) and for the coherent field transmission matrix (Eq. (12)); these are different objects and should be denoted separately (e.g., U for the field matrix).
  2. [§VI] In the S<N extension, 'forces each row to be a distinct canonical basis vector' should read 'each column'; in the S>N extension, the Gram matrix whose off-diagonals are forced to vanish is T T^T (N×N), not T^T T (S×S). Please correct the dimensions and wording.
  3. [Eq. (11)] The notation N_n and N_g is used without explicit definition; please define the ensemble sizes in the text or caption.
  4. [Eq. (4)] The high-SNR capacity formula is attributed to Ref. [50] and also to an SM derivation; since the SM is not included, please ensure the derivation is either provided or the attribution and conditions (peak vs. average power, Gaussian noise) are made explicit.

Circularity Check

0 steps flagged

No meaningful circularity: the permutation-optimality theorem is derived from the stated objectives, with only a non-circular omitted-SM gap for partially coherent inputs.

full rationale

The central derivation is self-contained. Section III states the objectives (Shannon surrogate in Eq. 5, Fisher surrogate in Eq. 9) and then proves, via the convexity argument on the singular-value simplex under the physical column-sum constraint, that the optimum has T^T T = I and hence, with nonnegative entries, T is a permutation matrix. The permutation conclusion is a consequence of the assumptions, not an input to them. The coherent-input proof of Sec. V is a separate argument: Eqs. (12)-(14) force each row of unitary T to have one nonzero entry, giving T = D_out Π D_in; this is not imported from prior work. The numerical examples optimize the stated objectives and reproduce the predicted permutation structure, so there is no fitted-input-called-prediction pattern. Self-citations are present (Ref. 17 by an author; Refs. 59, 60 by authors), but they are contextual: Ref. 17 motivates the desire for understanding, and Refs. 59, 60 support the generic singular-value decay used to justify regularization, not the permutation theorem. The one flagged issue is the partially-coherent extension in Sec. V, which is asserted by 'A complementary Fisher-information analysis confirms the same conclusion (cf. SM)'; the SM is not included, so that extension is unverified in the provided manuscript. That is an omitted-proof / completeness concern, not a circular reduction, and therefore does not raise the circularity score.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

The central theorem rests on standard matrix analysis (nonnegative orthonormal matrices are permutation matrices), the linear incoherent model, the passivity constraint, and the choice of noise model. No free parameters are fitted to data; the SNR regularization is physical. No new physical entities are introduced.

axioms (6)
  • domain assumption Incoherent imaging is linear: Y = T X (Eq. 1)
    The entire theoretical analysis is based on the linear PSF model, which requires mutually incoherent sources.
  • domain assumption T has nonnegative entries and column sums at most 1
    Nonnegativity from intensity-only detection; column-sum bound from passive power conservation. This is the key constraint in the convexity proof (Sec. III).
  • domain assumption Additive Gaussian readout noise (Eq. 2)
    The Shannon and Fisher objectives are derived under this noise model; the paper acknowledges shot-noise extension is future work.
  • domain assumption High-SNR capacity formula of Eq. (5) from Ref. [50]
    The channel capacity for intensity MIMO is imported from previous work; the paper uses it as an objective.
  • domain assumption For the coherent extension, T is unitary/lossless
    The proof that each row has at most one nonzero element requires T^† T = I (Sec. V).
  • domain assumption CRB applies: estimators are unbiased
    The Fisher objective E = σ^2 Tr[(T^T T)^{-1}] is the CRB for unbiased estimators; the paper acknowledges that biased/regularized estimators with priors can exceed this (Sec. VI).

pith-pipeline@v1.3.0-alltime-deepseek · 15564 in / 18253 out tokens · 195961 ms · 2026-08-02T11:09:32.130662+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of End-to-end meta-imagers: Information-theoretic objectives and generalized focusing optima." pith.science (2026). https://pith.science/paper/MRUUVRNN

@misc{pith2026260616724,
  author       = {Pith},
  title        = {Pith review of: End-to-end meta-imagers: Information-theoretic objectives and generalized focusing optima},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MRUUVRNN}},
  note         = {Machine review of arXiv:2606.16724}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Metasurfaces and complex photonic components are increasingly co-designed with computational back-ends via end-to-end optimization, yet such optimizations are expensive and opaque -- obscuring the role of the optics and any fundamental performance limits. We develop two information-theoretic objectives, based on Shannon capacity and Fisher information, that isolate the photonic contribution to image formation. Both are closed-form, data-free functions of the transfer matrix, requiring no training data, and yield designs whose reconstruction quality matches end-to-end optimization. We prove that for both objectives, and for a broader family with a shared mathematical structure, the optimal transfer matrix is a permutation matrix: each source's emission is concentrated on a single, distinct detector, a condition we call generalized focusing. This holds regardless of source/detector geometry, as we demonstrate in settings where conventional imaging intuition offers no guidance, including a two-way imager, imaging through a random scattering medium, and Hermite--Gauss mode sorting. The root of this constraint is an "intensity bottleneck": nonnegative intensity measurements admit only Kronecker deltas as a complete orthonormal basis. We further show that this bottleneck, and the generalized-focusing optimum, persist for coherent and partially coherent sources -- the constraint is the detector array, not the source coherence.

Figures

Figures reproduced from arXiv: 2606.16724 by Lukas Kienesberger, Owen D. Miller, Yaxi Liu, Zeyu Kuang.

Figure 1
Figure 1. Figure 1: FIG. 1: End-to-end computational imaging pipeline: sources ( [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2: Geometric interpretation of the two information-theoretic objectives, illustrated for [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: (a) Two-way imager with sources to its left and below (red) and detectors above and to its [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4: Imaging through a random scattering medium. (a) Schematic: sources (red) are embedded in [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5: Mode sorting of Hermite–Gauss beams. (a) Schematic: HG modes impinge from the left onto [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6: Coherent amplitude retrieval with intensity-only detection. (a) Schematic: coherent input [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

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

  1. Optically Incoherent Photonic Mutual Information

    physics.optics 2026-07 conditional novelty 7.0

    An end-to-end electromagnetic channel model shows point focusing is uniquely optimal for isotropic incoherent sources under a spectrum-flattening condition, while oversampled intensity detection favors interferometric...

  2. VLM-Aware Meta-Optic Front-End Design for Frozen Vision-Language Models

    cs.CV 2026-06 conditional novelty 6.5

    Optimizing a constrained meta-optic density for frozen CLIP cross-entropy, not focusing, raises ImageNet-100 zero-shot accuracy by about 12 points over a focus baseline and transfers across models and datasets.

  3. VLM-Aware Meta-Optic Front-End Design for Frozen Vision-Language Models

    cs.CV 2026-06 unverdicted novelty 6.0

    CODA optimizes continuous-density meta-optics via adjoint gradients on Maxwell simulations to boost frozen CLIP zero-shot accuracy on ImageNet-100 from 53.75% to 65.41%, with transfer to other models.

Reference graph

Works this paper leans on

83 extracted references · 2 canonical work pages · cited by 2 Pith papers · 1 internal anchor

  1. [1]

    J. N. Mait, G. W. Euliss, and R. A. Athale, Computational imaging, Adv. Opt. Photon.10, 409 (2018)

    2018

  2. [2]

    Delbracio, D

    M. Delbracio, D. Kelly, M. S. Brown, and P. Milanfar, Mobile computational photography: A tour, Annual Review of Vision Science7, 571 (2021)

    2021

  3. [3]

    Yu and F

    N. Yu and F. Capasso, Flat optics with designer metasurfaces, Nature Materials13, 139 (2014)

    2014

  4. [4]

    Khorasaninejad, W

    M. Khorasaninejad, W. T. Chen, R. C. Devlin, J. Oh, A. Y. Zhu, and F. Capasso, Metalenses at visible wavelengths: Diffraction-limited focusing and subwavelength resolution imaging, Science352, 1190 (2016)

    2016

  5. [5]

    W. T. Chen, A. Y. Zhu, and F. Capasso, Flat optics with dispersion-engineered metasurfaces, Nature Reviews Materials5, 604 (2020)

    2020

  6. [6]

    Rotter and S

    S. Rotter and S. Gigan, Light fields in complex media: Mesoscopic scattering meets wave control, Rev. Mod. Phys.89, 015005 (2017)

    2017

  7. [7]

    D. G. Stork and M. D. Robinson, Theoretical foundations for joint digital-optical analysis of electro-optical imaging systems, Appl. Opt.47, B64 (2008)

    2008

  8. [8]

    Chakrabarti, Learning sensor multiplexing design through back-propagation, inAdvances in Neural Information Processing Systems (NeurIPS), Vol

    A. Chakrabarti, Learning sensor multiplexing design through back-propagation, inAdvances in Neural Information Processing Systems (NeurIPS), Vol. 29 (2016) pp. 3081–3089, arXiv:1605.07078

    Pith/arXiv arXiv 2016

  9. [9]

    Sitzmann, S

    V. Sitzmann, S. Diamond, Y. Peng, X. Dun, S. Boyd, W. Heidrich, F. Heide, and G. Wetzstein, End-to-end optimization of optics and image processing for achromatic extended depth of field and super-resolution imaging, ACM Trans. Graph.37, 10.1145/3197517.3201333 (2018)

    arXiv 2018

  10. [10]

    Chang, V

    J. Chang, V. Sitzmann, X. Dun, W. Heidrich, and G. Wetzstein, Hybrid optical-electronic convolutional neural networks with optimized diffractive optics for image classification, Scientific Reports8, 12324 (2018)

    2018

  11. [11]

    Tseng, S

    E. Tseng, S. Colburn, J. Whitehead, L. Huang, S.-H. Baek, A. Majumdar, and F. Heide, Neural nano- optics for high-quality thin lens imaging, Nature Communications12, 6493 (2021)

    2021

  12. [12]

    Z. Lin, C. Roques-Carmes, R. Pestourie, M. Soljaˇ ci´ c, A. Majumdar, and S. G. Johnson, End-to-end nanophotonic inverse design for imaging and polarimetry, Nanophotonics10, 1177 (2021)

    2021

  13. [13]

    Z. Lin, R. Pestourie, C. Roques-Carmes, Z. Li, F. Capasso, M. Soljaˇ ci´ c, and S. G. Johnson, End-to-end metasurface inverse design for single-shot multi-channel imaging, Opt. Express30, 28358 (2022)

    2022

  14. [14]

    G. Arya, W. F. Li, C. Roques-Carmes, M. Soljaˇ ci´ c, S. G. Johnson, and Z. Lin, End-to-end optimization of metasurfaces for imaging with compressed sensing, ACS Photonics11, 2077 (2024)

    2077

  15. [15]

    Wetzstein, A

    G. Wetzstein, A. Ozcan, S. Gigan, S. Fan, D. Englund, M. Soljaˇ ci´ c, C. Denz, D. A. B. Miller, and D. Psaltis, Inference in artificial intelligence with deep optics and photonics, Nature588, 39 (2020)

    2020

  16. [16]

    Barbastathis, A

    G. Barbastathis, A. Ozcan, and G. Situ, On the use of deep learning for computational imaging, Optica 6, 921 (2019)

    2019

  17. [17]

    The hidden dimension in nanophotonics design: understanding

    P. Lalanne and O. Miller, The hidden dimension in nanophotonics design: understanding, arXiv [physics.optics] 10.48550/arXiv.2604.07860 (2026), arXiv:2604.07860 [physics.optics]

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2604.07860 2026

  18. [18]

    E. E. Fenimore and T. M. Cannon, Coded aperture imaging with uniformly redundant arrays, Appl. Opt. 17, 337 (1978)

    1978

  19. [19]

    Levin, R

    A. Levin, R. Fergus, F. Durand, and W. T. Freeman, Image and depth from a conventional camera with a coded aperture, ACM Trans. Graph.26, 10.1145/1276377.1276464 (2007)

    arXiv 2007

  20. [20]

    E. R. Dowski and W. T. Cathey, Extended depth of field through wave-front coding, Appl. Opt.34, 1859 (1995)

    1995

  21. [21]

    S. C. Park, M. K. Park, and M. G. Kang, Super-resolution image reconstruction: a technical overview, IEEE Signal Processing Magazine20, 21 (2003)

    2003

  22. [22]

    J. Yang, J. Wright, T. S. Huang, and Y. Ma, Image super-resolution via sparse representation, IEEE Transactions on Image Processing19, 2861 (2010). 13

    2010

  23. [23]

    C. Dong, C. C. Loy, K. He, and X. Tang, Image super-resolution using deep convolutional networks, IEEE Transactions on Pattern Analysis and Machine Intelligence38, 295 (2016)

    2016

  24. [24]

    Colburn, A

    S. Colburn, A. Zhan, and A. Majumdar, Metasurface optics for full-color computational imaging, Science Advances4, eaar2114 (2018)

    2018

  25. [25]

    Bayati, R

    E. Bayati, R. Pestourie, S. Colburn, Z. Lin, S. G. Johnson, and A. Majumdar, Inverse designed extended depth of focus meta-optics for broadband imaging in the visible, Nanophotonics11, 2531 (2022)

    2022

  26. [26]

    Chakravarthula, J

    P. Chakravarthula, J. Sun, X. Li, C. Lei, G. Chou, M. Bijelic, J. Froech, A. Majumdar, and F. Heide, Thin on-sensor nanophotonic array cameras, ACM Trans. Graph.42, 10.1145/3618398 (2023)

    work page doi:10.1145/3618398 2023

  27. [27]

    P. B. Fellgett and E. H. Linfoot, On the assessment of optical images, Philosophical Transactions of the Royal Society of London, Series A: Mathematical and Physical Sciences247, 369 (1955)

    1955

  28. [28]

    E. H. Linfoot, Information theory and optical images, J. Opt. Soc. Am.45, 808 (1955)

    1955

  29. [29]

    Toraldo di Francia, Degrees of freedom of an image, J

    G. Toraldo di Francia, Degrees of freedom of an image, J. Opt. Soc. Am.59, 799 (1969)

    1969

  30. [30]

    F. O. Huck, C. L. Fales, R. Alter-Gartenberg, S. K. Park, and Z.-u. Rahman, Information-theoretic assessment of sampled imaging systems, Optical Engineering38, 742 (1999)

    1999

  31. [31]

    Ashok and M

    A. Ashok and M. A. Neifeld, Information-based analysis of simple incoherent imaging systems, Opt. Express11, 2153 (2003)

    2003

  32. [32]

    D. A. B. Miller, Waves, modes, communications, and optics: a tutorial, Adv. Opt. Photon.11, 679 (2019)

    2019

  33. [33]

    Pinkard, L

    H. Pinkard, L. Kabuli, E. Markley, T. Chien, J. Jiao, and L. Waller, Information-driven design of imaging systems (2024), arXiv:2405.20559 [physics.optics]

    arXiv 2024

  34. [34]

    Markley, H

    E. Markley, H. Pinkard, L. Kabuli, N. Singh, and L. Waller, Computationally efficient information-driven optical design with interchanging optimization (2025), arXiv:2507.07789 [eess.IV]

    Pith/arXiv arXiv 2025

  35. [35]

    L. A. Kabuli, H. Pinkard, E. Markley, C. S. Hung, and L. Waller, Designing lensless imaging systems to maximize information capture, Optica13, 227 (2026)

    2026

  36. [36]

    Tsang, R

    M. Tsang, R. Nair, and X.-M. Lu, Quantum theory of superresolution for two incoherent optical point sources, Phys. Rev. X6, 031033 (2016)

    2016

  37. [37]

    Nair and M

    R. Nair and M. Tsang, Far-field superresolution of thermal electromagnetic sources at the quantum limit, Phys. Rev. Lett.117, 190801 (2016)

    2016

  38. [38]

    Lupo and S

    C. Lupo and S. Pirandola, Ultimate precision bound of quantum and subwavelength imaging, Phys. Rev. Lett.117, 190802 (2016)

    2016

  39. [39]

    Pa´ ur, B

    M. Pa´ ur, B. Stoklasa, Z. Hradil, L. L. S´ anchez-Soto, and J. Re ˇh´ aˇ cek, Achieving the ultimate optical resolution, Optica3, 1144 (2016)

    2016

  40. [40]

    W.-K. Tham, H. Ferretti, and A. M. Steinberg, Beating Rayleigh’s curse by imaging using phase informa- tion, Phys. Rev. Lett.118, 070801 (2017)

    2017

  41. [41]

    Tsang, Resolving starlight: a quantum perspective, Contemporary Physics60, 279 (2019)

    M. Tsang, Resolving starlight: a quantum perspective, Contemporary Physics60, 279 (2019)

    2019

  42. [42]

    S. A. Wadood, S. Aarav, K. Liang, and J. W. Fleischer, Super-resolution with Fourier measurements (2025), arXiv:2511.06098 [physics.optics]

    arXiv 2025

  43. [43]

    J. R. Janesick,Photon Transfer, SPIE Press Monograph, Vol. PM170 (SPIE Press, Bellingham, W A, 2007)

    2007

  44. [44]

    G. E. Healey and R. Kondepudy, Radiometric CCD camera calibration and noise estimation, IEEE Trans- actions on Pattern Analysis and Machine Intelligence16, 267 (1994)

    1994

  45. [45]

    T. M. Cover and J. A. Thomas,Elements of Information Theory, 2nd ed. (Wiley-Interscience, Hoboken, NJ, 2006)

    2006

  46. [46]

    Telatar, Capacity of multi-antenna gaussian channels, European Transactions on Telecommunications 10, 585 (1999)

    E. Telatar, Capacity of multi-antenna gaussian channels, European Transactions on Telecommunications 10, 585 (1999)

    1999

  47. [47]

    G. J. Foschini and M. J. Gans, On limits of wireless communications in a fading environment when using multiple antennas, Wireless Personal Communications6, 311 (1998)

    1998

  48. [48]

    Lapidoth, S

    A. Lapidoth, S. M. Moser, and M. A. Wigger, On the capacity of free-space optical intensity channels, IEEE Transactions on Information Theory55, 4449 (2009)

    2009

  49. [49]

    L. Li, S. M. Moser, L. Wang, and M. Wigger, On the capacity of mimo optical wireless channels, IEEE Transactions on Information Theory66, 5660 (2020)

    2020

  50. [50]

    S. M. Moser, M. Mylonakis, L. Wang, and M. Wigger, Asymptotic capacity results for MIMO wireless optical communication, in2017 IEEE International Symposium on Information Theory (ISIT)(2017) pp. 536–540

    2017

  51. [51]

    H. H. Barrett and K. J. Myers,Foundations of Image Science, Wiley Series in Pure and Applied Optics (Wiley-Interscience, Hoboken, NJ, 2004)

    2004

  52. [52]

    R. E. Thompson, D. R. Larson, and W. W. Webb, Precise nanometer localization analysis for individual fluorescent probes, Biophysical Journal82, 2775 (2002)

    2002

  53. [53]

    S. Ram, E. S. Ward, and R. J. Ober, Beyond Rayleigh’s criterion: A resolution measure with application to single-molecule microscopy, Proc. Natl. Acad. Sci. USA103, 4457 (2006)

    2006

  54. [54]

    J. Chao, E. S. Ward, and R. J. Ober, Fisher information theory for parameter estimation in single molecule microscopy: tutorial, J. Opt. Soc. Am. A33, B36 (2016)

    2016

  55. [55]

    S. R. P. Pavani, M. A. Thompson, J. S. Biteen, S. J. Lord, N. Liu, R. J. Twieg, R. Piestun, and W. E. Moerner, Three-dimensional, single-molecule fluorescence imaging beyond the diffraction limit by using a double-helix point spread function, Proc. Natl. Acad. Sci. USA106, 2995 (2009). 14

    2009

  56. [56]

    Shechtman, S

    Y. Shechtman, S. J. Sahl, A. S. Backer, and W. E. Moerner, Optimal point spread function design for 3d imaging, Phys. Rev. Lett.113, 133902 (2014)

    2014

  57. [57]

    J. Yu, H. Lo, W. Chen, C. Zhu, Y. Wu, F. Wang, C. Wang, C. Yan, C. Dang, B. Wen, H. Cao, Y. Chong, and Q. J. Wang, Wavelength-scale noise-resistant on-chip spectrometer, arXiv preprint arXiv:2509.22286 (2025), arXiv:2509.22286

    arXiv 2025

  58. [58]

    W. Ma, R. Pestourie, Z. Lin, and S. G. Johnson, Inverse design for robust inference in integrated compu- tational spectrometry, Nanophotonics15, e70054 (2026)

    2026

  59. [59]

    Kuang, D

    Z. Kuang, D. A. B. Miller, and O. D. Miller, Bounds on the coupling strengths of communication channels and their information capacities, IEEE Transactions on Antennas and Propagation73, 3959 (2025)

    2025

  60. [60]

    D. A. B. Miller, Z. Kuang, and O. D. Miller, Tunneling escape of waves, Nature Photonics19, 284 (2025)

    2025

  61. [61]

    Morizur, L

    J.-F. Morizur, L. Nicholls, P. Jian, S. Armstrong, N. Treps, B. Hage, M. Hsu, W. Bowen, J. Janousek, and H.-A. Bachor, Programmable unitary spatial mode manipulation, J. Opt. Soc. Am. A27, 2524 (2010)

    2010

  62. [62]

    Labroille, B

    G. Labroille, B. Denolle, P. Jian, P. Genevaux, N. Treps, and J.-F. Morizur, Efficient and mode selective spatial mode multiplexer based on multi-plane light conversion, Opt. Express22, 15599 (2014)

    2014

  63. [63]

    N. K. Fontaine, R. Ryf, H. Chen, D. T. Neilson, K. Kim, and J. Carpenter, Laguerre-Gaussian mode sorter, Nature Communications10, 1865 (2019)

    2019

  64. [64]

    D. A. B. Miller, Self-configuring universal linear optical component, Photonics Research1, 1 (2013)

    2013

  65. [65]

    I. M. Vellekoop and A. P. Mosk, Focusing coherent light through opaque strongly scattering media, Opt. Lett.32, 2309 (2007)

    2007

  66. [66]

    S. M. Popoff, G. Lerosey, R. Carminati, M. Fink, A. C. Boccara, and S. Gigan, Measuring the transmission matrix in optics: An approach to the study and control of light propagation in disordered media, Phys. Rev. Lett.104, 100601 (2010)

    2010

  67. [67]

    A. P. Mosk, A. Lagendijk, G. Lerosey, and M. Fink, Controlling waves in space and time for imaging and focusing in complex media, Nature Photonics6, 283 (2012)

    2012

  68. [68]

    H. Cao, A. P. Mosk, and S. Rotter, Shaping the propagation of light in complex media, Nature Physics 18, 994 (2022)

    2022

  69. [69]

    Bouchet, S

    D. Bouchet, S. Rotter, and A. P. Mosk, Maximum information states for coherent scattering measurements, Nature Physics17, 564 (2021)

    2021

  70. [70]

    H¨ upfl, F

    J. H¨ upfl, F. Russo, L. M. Rachbauer, D. Bouchet, J. Lu, U. Kuhl, and S. Rotter, Continuity equation for the flow of Fisher information in wave scattering, Nature Physics20, 1294 (2024)

    2024

  71. [71]

    Starshynov, M

    I. Starshynov, M. Weimar, L. M. Rachbauer, G. Hackl, D. Faccio, S. Rotter, and D. Bouchet, Model-free estimation of the Cram´ er–Rao bound for deep learning microscopy in complex media, Nature Photonics 19, 593 (2025)

    2025

  72. [72]

    Q. Sun, C. Wang, Q. Fu, X. Dun, and W. Heidrich, End-to-end complex lens design with differentiable ray tracing, ACM Trans. Graph.40, 10.1145/3450626.3459674 (2021)

    arXiv 2021

  73. [73]

    J. R. Fienup, Phase retrieval algorithms: a comparison, Applied Optics21, 2758 (1982)

    1982

  74. [74]

    Shechtman, Y

    Y. Shechtman, Y. C. Eldar, O. Cohen, H. N. Chapman, J. Miao, and M. Segev, Phase retrieval with application to optical imaging: a contemporary overview, IEEE Signal Processing Magazine32, 87 (2015)

    2015

  75. [75]

    E. H. Adelson and J. Y. Wang, Single lens stereo with a plenoptic camera, IEEE Transactions on Pattern Analysis and Machine Intelligence14, 99 (1992)

    1992

  76. [76]

    Ng,Digital light field photography, Ph.D

    R. Ng,Digital light field photography, Ph.D. thesis, Stanford University (2006)

    2006

  77. [77]

    Lustig, D

    M. Lustig, D. Donoho, and J. M. Pauly, Sparse MRI: The application of compressed sensing for rapid MR imaging, Magnetic Resonance in Medicine58, 1182 (2007)

    2007

  78. [78]

    D. L. Donoho and I. M. Johnstone, Ideal spatial adaptation by wavelet shrinkage, Biometrika81, 425 (1994)

    1994

  79. [79]

    E. J. Cand` es, J. Romberg, and T. Tao, Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information, IEEE Transactions on Information Theory52, 489 (2006)

    2006

  80. [80]

    Keshvari, W

    A. Keshvari, W. Tuxbury, and Z. Lin, Adaptive sensing beyond non-adaptive information limits: End-to-end co-design of geometry, policy, and inference, arXiv preprint arXiv:2604.25193 (2026), arXiv:2604.25193

    Pith/arXiv arXiv 2026

Showing first 80 references.