Pith. sign in

REVIEW 5 major objections 8 minor 57 references

SH-SAS: An Implicit Neural Representation for Complex Spherical-Harmonic Scattering Fields for 3D Synthetic Aperture Sonar

T0 review · 5 major / 8 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read Parameterizing acoustic scattering as spherical-harmonic coefficients inside a neural field improves 3D sonar reconstruction from raw 1D time-of-flight signals.

desk verdict Useful representation for directional SAS, but the forward model breaks reciprocity and the synthetic validation is circular; real claims rest on qualitative images. read the letter →

arxiv 2509.11087 v1 pith:MYMKYBCE submitted 2025-09-14 cs.GR cs.CVcs.LG

Omkar Shailendra Vengurlekar , Adithya Pediredla , Suren Jayasuriya This is my paper · ORCID
classification cs.GRcs.CVcs.LG
keywords syntheticaperturesonarimplicitneuralrepresentationsphericalharmonicsacousticscatteringfieldtime-of-flight3Dreconstructionrenderinganisotropic
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

SH-SAS aims to show that replacing the isotropic scattering-density assumption in neural volumetric sonar reconstruction with a spherical-harmonic (SH) expansion of the complex scattering field produces higher-fidelity 3D reconstructions. The DC term of the SH expansion supplies a stable density proxy, while higher-order coefficients capture direction-dependent scattering with minimal parameter cost. Training runs directly on 1D time-of-flight signals through a differentiable forward model, so no beamformed intermediate images are needed. On simulated, in-air, and underwater benchmarks, the method reports lower Chamfer distance and better IoU/precision/F1 than previous isotropic neural methods and time-domain backprojection.

What carries the argument

The central object is the spherical-harmonic expansion of the scattering field, σ_s(x,θ,ϕ)=Σ_{l=0}^{L}Σ_{m=-l}^{l} c_{l,m}(x)Y_{l,m}(θ,ϕ), where Y_{l,m} are the spherical-harmonic basis functions on the unit sphere and c_{l,m}(x) are complex learnable coefficients. Spherical harmonics are an orthonormal angular basis, so a small set of coefficients captures smoothly varying directionality. The zeroth-order coefficient supplies the isotropic density and surface normals, while higher orders encode anisotropic scattering. The forward model integrates these values over constant-time-of-flight ellipsoids with transmitter/receiver directivity and a Lambertian cosine incident-angle term, synthesizi

What would settle it

A controlled experiment with a flat, smooth target at a grazing angle, where a strong specular return produces a joint transmit/receive direction dependence that the product of an SH receive function and a Lambertian incident cosine cannot fit, would falsify the central claim if the measured transient residuals remain large at high SNR even as SH degree L increases. More simply: if increasing L from 1 to 3 never improves Chamfer distance on any benchmark, then the directionality modeling is not carrying the improvement.

Watch

Extended reading notes

Core claim

The central claim is that the complex acoustic scattering field at each spatial point can be compactly represented by complex spherical-harmonic coefficients up to degree L, predicted by a lightweight hash-encoded MLP, and that optimizing this field directly against 1D time-of-flight signals yields 3D synthetic aperture sonar reconstructions with sharper geometry and fewer artifacts than isotropic neural volumetric baselines or time-domain backprojection. The zeroth-order SH coefficient defines the isotropic density, whose gradient gives surface normals; higher orders add angular resolution, which the paper argues is the inductive bias that improves reconstruction under occlusions, sparse ap

Load-bearing premise

The load-bearing premise is that the true acoustic scattering at each point factors into a spherical-harmonic function of the receiver-to-point direction times a Lambertian cosine of the transmitter-to-point direction; scattering that depends on both directions jointly—such as specular reflection or diffraction—cannot be captured, and the paper explicitly excludes diffraction and sub-wavelength wave effects from its forward model.

Editorial extensions

If this is right

  • Higher SH degrees (up to L=3) monotonically improve reconstruction quality in the paper's ablation, indicating that directionality, not just density, is what the network learns.
  • Because training uses raw 1D ToF signals, the same architecture could be applied to other coherent time-of-flight sensors by swapping the forward model.
  • The representation handles sparse aperture subsampling (20% of views) with less streaking and fragmentation than backprojection or the isotropic neural baseline.
  • The learned field is stable under iso-surface threshold changes, meaning mesh extraction is less sensitive to an arbitrary threshold.
  • On a single-look underwater acquisition, the method recovers the expected one-sided flat profile and suppresses returns beneath the target, consistent with occluded geometry.

Reading between the lines

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

  • Inference: If the factorized SH-Lambertian scattering model is accurate, the same parameterization could benefit radar and LiDAR neural reconstruction, where view-dependent coherent scattering also matters; the paper lists this as future work.
  • Inference: A quantitative test of the directionality claim would compare SH-SAS with the isotropic baseline on targets with known specular or strongly bistatic responses, where the factorized model should be wrong; the paper's real-data experiments are primarily qualitative.
  • Inference: The stated sensitivity of higher-order SH terms to noise suggests that an adaptive degree per spatial region, or a spectral prior on the SH coefficients, could extend the method to low-SNR and narrowband acquisitions.
  • Inference: Since the DC term alone reproduces the isotropic baseline, the practical gain of the method could be measured as a function of view diversity; one would expect the largest improvement in sparse-aperture regimes, consistent with the 20%-views experiment.
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.

Referee Report

5 major / 8 minor

Summary. The paper proposes SH-SAS, an implicit neural representation for 3D synthetic aperture sonar reconstruction. Instead of modeling each voxel as an isotropic complex scattering density as in Reed et al., SH-SAS outputs spherical harmonic (SH) coefficients up to degree L at each point; the DC term provides an isotropic density proxy and higher-order terms encode directional scattering. Training is end-to-end from 1D time-of-flight signals through a differentiable point-based forward model with ellipsoidal sampling and pulse deconvolution. Experiments on synthetic scenes, in-air AirSAS data, and underwater SVSS data report improved reconstruction quality (Chamfer distance, IoU, precision, F1) over time-domain backprojection and Reed et al., along with faster convergence and better performance under sparse views.

Significance. If the claims hold, this is a useful step for SAS: replacing isotropic scattering with a compact directional SH representation inside a neural volumetric inversion, using a lightweight hash-encoded MLP and training directly from raw ToF signals. The paper includes a credible synthetic evaluation, a thoughtful ablation of SH levels, and a novel-view synthesis experiment. However, the evidence has important gaps: the synthetic simulator's scattering law is not specified, the forward model's factorization breaks acoustic reciprocity, real-data results are only qualitative, and per-scene synthetic metrics are not uniformly better than the baseline. These issues must be addressed before the contribution can be fully assessed.

major comments (5)
  1. [Sec. 4.4, Eq. (6)] The factorized forward model sigma_s(x, theta, phi) * g(omega_x) is not invariant under exchanging TX and RX, violating acoustic reciprocity for passive linear media. The model class therefore excludes physically reciprocal scattering functions. The manuscript does not discuss or justify this approximation. Please either provide a physical justification, demonstrate empirically that the learned field is approximately reciprocal (e.g., by comparing predictions for swapped TX/RX on a trained scene), or explicitly state this as a limitation affecting the physical interpretability of the SH coefficients.
  2. [Sec. 5.1] The synthetic data generator is described only as 'geometric acoustics' with diffraction and sub-wavelength effects excluded; the scattering law is not stated. If the simulator uses the same factorized model as the reconstruction, the synthetic experiments mainly test inversion self-consistency rather than the ability to capture true bistatic directionality. If it uses a reciprocal scattering law, the data lie outside the model class and the gains over the isotropic baseline need explanation. Please specify the scattering law used in the simulator and, if possible, run a test where the simulator includes a genuinely bidirectional (e.g., specular or angle-dependent) response to verify that the SH terms capture it.
  3. [Secs. 6.2 and 6.6] Real-data evaluations (AirSAS and SVSS) are qualitative only. Since AirSAS uses known 3D-printed objects, quantitative metrics (e.g., Chamfer distance or IoU against ground-truth meshes) should be reported. The SVSS result is validated only against expected geometry. Without quantitative real-data evidence, the claim that SH-SAS generalizes to real SAS is not strongly supported. Please add metrics for at least the AirSAS objects and discuss the SVSS case in terms of measurable criteria where possible.
  4. [Sec. 6.1 and Table 3] The average metrics in Table 1 hide per-scene variability. In Table 3, Reed et al. achieves better or comparable IoU/F1 on Bunny and Dragon point clouds (e.g., Bunny IoU 0.412 vs 0.398; Dragon F1 0.529 vs 0.526). The blanket claim of 'better performance' is therefore not uniformly supported. Report per-scene results with error bars (multiple runs or seeds) and discuss the cases where the SH model does not improve over the isotropic baseline.
  5. [Sec. 5.4 and Eq. (8)] Key hyperparameters are not reported: the occlusion scaling factor zeta in Eq. (4) and the loss weights lambda1-lambda5 in Eq. (8). The text also states that priors are 'disabled for the Simulated and SVSS datasets' without clarifying whether this applies to both SH-SAS and Reed et al. This ambiguity affects the fairness of the comparison and the reproducibility of the results. Please provide all hyperparameter values and state explicitly which regularization terms are used for each method and dataset.
minor comments (8)
  1. [Fig. 3 caption] Typo: 'Armadilo' should be 'Armadillo'.
  2. [Supplement Sec. 9] In the ellipsoid definition, 'b=c= sqrt(a^2 - (d/2)^2)' is confusing: c is already used for sound speed. It should be b = sqrt(a^2 - (d/2)^2).
  3. [Eq. (6)] State explicitly whether the integral is over a volumetric shell or a surface and clarify the meaning of dx.
  4. [Sec. 4.2] Specify the normalization and convention for the spherical harmonics (real vs complex, Condon-Shortley phase) used in the implementation.
  5. [Sec. 5.4] The transmitter beam pattern b_T(x) is never defined; its value/functional form should be given for each dataset, as it affects the forward model.
  6. [Eq. (8)] LScatterTV applies total variation to the complex field sigma_s; clarify whether it acts on the magnitude, phase, or complex values.
  7. [Sec. 5.4] The paper states code will be released after peer review but also includes code in the supplement; align these statements to avoid confusion.
  8. [Sec. 6.5] The SH-level ablation is reported only qualitatively. A quantitative table (e.g., Chamfer/IoU for L=1,2,3) would strengthen the claim that higher SH degrees are beneficial.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; forward model reuse in synthetic experiments is a self-consistency check, and real-data experiments provide independent support.

full rationale

The paper's derivation chain is an analysis-by-synthesis inversion: the forward model in Eq. (6) is inherited from Reed et al. (a published prior work, with two shared authors), and the SH representation in Eq. (2) is the paper's addition, trained by comparing synthesized ToF signals against measured or pulse-deconvolved signals. No step defines the target reconstruction in terms of the input measurements by construction. The synthetic data generator (Sec. 5.1) is described only as a 'transient time-of-flight (ToF) renderer' using 'geometric acoustics,' and it is not explicitly stated to use the same SH-parameterized forward model as the reconstruction, so one cannot exhibit a specific reduction of Eq. (6) to the simulator. Even if the simulator did use the same forward model, that would be a self-consistency test of inversion rather than a validation of the scattering physics, and the paper additionally reports real AirSAS and SVSS results and held-out novel-view synthesis, which are external benchmarks. The citations to Reed et al. are for a physics-based forward model and pulse-deconvolution scheme, not for an unverified uniqueness claim, so they are not load-bearing in a circular sense. The Limitations section honestly states that diffraction, interference, and multiple scattering are not modeled; this is a modeling-scope issue, not a circularity. Therefore, no circular step is identifiable with the required textual evidence.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The reconstruction rests on a point-based, single-scattering forward model inherited from prior work by the same group; the main new assumption is the factorization of directionality into receive-direction SH and an incident Lambertian cosine, plus the unstated zeta and loss weights.

free parameters (4)
  • SH degree L = 3
    Maximum harmonic order chosen by hand; all experiments use L=3 (Section 5.4).
  • occlusion scaling factor zeta = not reported
    Density is set as |sigma_s,DC|^zeta in Eq. (4); zeta is never given a value in the paper.
  • loss weights lambda1-lambda5 = not reported
    Composite loss in Eq. (8) has five weighting terms; their numeric values are not stated.
  • pulse deconvolution weights lambda1, lambda2 = not reported
    In Eq. (1), LPD regularization weights are not specified, affecting preprocessing quality.
assumptions (6)
  • domain assumption Point-based scattering model with complex analytic signals (Eq. 6)
    Borrowed from Brown et al. and Reed et al.; treats each point as an independent scatterer with a complex amplitude, no interactions between scatterers.
  • domain assumption Lambertian cosine g(omega_x) for incident angle dependence
    The scattering amplitude is multiplied by max(0, n.(o_T-x)/||o_T-x||), a diffuse scattering assumption.
  • domain assumption SH representation is band-limited to L=3
    Assumes directional scattering has limited angular bandwidth; higher L not tested in main results.
  • ad hoc to paper Density from DC SH magnitude: rho = |sigma_s,DC|^zeta
    The isotropic DC component is used as a geometric occupancy proxy; zeta is unstated.
  • standard math Transmission probability follows Beer-Lambert accumulation (Eq. 7)
    Standard volume rendering transmittance used in NeRF-style models.
  • domain assumption Ellipsoidal sampling restricts integration to constant-ToF ellipsoid
    Assumes the measured ToF bin only receives contributions from points on the corresponding ellipsoid, ignoring multipath.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SH-SAS: An Implicit Neural Representation for Complex Spherical-Harmonic Scattering Fields for 3D Synthetic Aperture Sonar." pith.science (2026). https://pith.science/paper/MYMKYBCE

@misc{pith2026250911087,
  author       = {Pith},
  title        = {Pith review of: SH-SAS: An Implicit Neural Representation for Complex Spherical-Harmonic Scattering Fields for 3D Synthetic Aperture Sonar},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MYMKYBCE}},
  note         = {Machine review of arXiv:2509.11087}
}
read the original abstract

Synthetic aperture sonar (SAS) reconstruction requires recovering both the spatial distribution of acoustic scatterers and their direction-dependent response. Time-domain backprojection is the most common 3D SAS reconstruction algorithm, but it does not model directionality and can suffer from sampling limitations, aliasing, and occlusion. Prior neural volumetric methods applied to synthetic aperture sonar treat each voxel as an isotropic scattering density, not modeling anisotropic returns. We introduce SH-SAS, an implicit neural representation that expresses the complex acoustic scattering field as a set of spherical harmonic (SH) coefficients. A multi-resolution hash encoder feeds a lightweight MLP that outputs complex SH coefficients up to a specified degree L. The zeroth-order coefficient acts as an isotropic scattering field, which also serves as the density term, while higher orders compactly capture directional scattering with minimal parameter overhead. Because the model predicts the complex amplitude for any transmit-receive baseline, training is performed directly from 1-D time-of-flight signals without the need to beamform intermediate images for supervision. Across synthetic and real SAS (both in-air and underwater) benchmarks, results show that SH-SAS performs better in terms of 3D reconstruction quality and geometric metrics than previous methods.

Figures

Figures reproduced from arXiv: 2509.11087 by the authors.

Figure 1
Figure 1. Pipeline: Neural-field architecture for Synthetic Aperture Sonar Reconstruction. Our approach uses hash encodings and a multi [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. Real AirSAS results for Armadilo and Bunny at [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 2
Figure 2. Simulated results for Buddha (top row) and Armadillo [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Sparse subsampling with 20% of views. Top: front view. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Convergence comparison on XYZ Dragon. Our method [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Underwater reconstruction from SVSS. Compared to backprojection and Reed et al. [ [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Increase in Spherical Harmonic Levels. As the Spherical Harmonic (SH) level increases from Level 1 to Level 3, the quality of [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Mesh reconstructions of synthetic data: Armadillo, Bunny, Happy Buddha, XYZ Dragon (side, top) scenes using (from left to [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Mesh consistency across threshold levels. Columns: methods. Rows: increasing threshold [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Real data: Armadillo. Top: frontal view. Bottom: side view. [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]
Figure 11
Figure 11. Figure 11: Real data: Stanford bunny. Top: frontal view. Bottom: top view. [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]
Figure 12
Figure 12. Figure 12: Novel view signal synthesis results comparing Reed et al. [ [PITH_FULL_IMAGE:figures/full_fig_p017_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 2 linked inside Pith

  1. [1]

    T ¨orf: Time-of-flight radiance fields for dynamic scene view synthesis.Advances in Neural Information Pro- cessing Systems, 34, 2021

    Benjamin Attal, Eliot Laidlaw, Aaron Gokaslan, Changil Kim, Christian Richardt, James Tompkin, and Matthew O’Toole. T ¨orf: Time-of-flight radiance fields for dynamic scene view synthesis.Advances in Neural Information Pro- cessing Systems, 34, 2021. 2

    2021

  2. [2]

    Gpu-based real-time synthetic aperture sonar processing on-board autonomous underwater vehi- cles

    Francesco Baralli, Michel Couillard, Jesus Ortiz, and Dar- win G Caldwell. Gpu-based real-time synthetic aperture sonar processing on-board autonomous underwater vehi- cles. In2013 MTS/IEEE OCEANS-Bergen, pages 1–8. IEEE,

  3. [3]

    Multi-view 3d surface reconstruction from sar images by inverse rendering.arXiv preprint arXiv:2502.10492, 2025

    Emile Barbier-Renard, Florence Tupin, Nicolas Trouv ´e, and Lo ¨ıc Denis. Multi-view 3d surface reconstruction from sar images by inverse rendering.arXiv preprint arXiv:2502.10492, 2025. 2

    arXiv 2025

  4. [4]

    Design and experimental results of a 300-khz synthetic aperture sonar optimized for shallow-water operations.IEEE Journal of Oceanic Engi- neering, 34(3):285–293, 2008

    Andrea Bellettini and Marc Pinto. Design and experimental results of a 300-khz synthetic aperture sonar optimized for shallow-water operations.IEEE Journal of Oceanic Engi- neering, 34(3):285–293, 2008. 1

    2008

  5. [5]

    Blanford, John D

    Thomas E. Blanford, John D. McKay, Daniel C. Brown, Joonho D. Park, and Shawn F. Johnson. Development of an in-air circular synthetic aperture sonar system as an edu- cational tool.Proceedings of Meetings on Acoustics, 36(1): 070002, 2019. 5

    2019

  6. [6]

    Leveraging audio hardware for under- water acoustics experiments

    Thomas E Blanford, Luke Garrett, J Daniel Park, and Daniel C Brown. Leveraging audio hardware for under- water acoustics experiments. InProceedings of Meetings on Acoustics 182ASA, page 030002. Acoustical Society of America, 2022. 5

    2022

  7. [7]

    An in-air synthetic aperture sonar dataset of target scattering in environments of varying complexity

    Thomas E Blanford, David P Williams, J Daniel Park, Brian T Reinhardt, Kyle S Dalton, Shawn F Johnson, and Daniel C Brown. An in-air synthetic aperture sonar dataset of target scattering in environments of varying complexity. Scientific Data, 11(1):1196, 2024. 1

    2024

  8. [8]

    Borts, E

    D. Borts, E. Liang, T. Broedermann, A. Ramazzina, S. Walz, E. Palladin, J. Sun, D. Brueggemann, C. Sakaridis, L. Van Gool, M. Bijelic, and F. Heide. Radar fields: Frequency- space neural scene representations for fmcw radar. InProc. SIGGRAPH, 2024. 2

    2024

Show all 57 references
  1. [9]

    PhD the- sis, The Pennsylvania State University, 2017

    Daniel C Brown.Modeling and measurement of spatial co- herence for normal incidence seafloor scattering. PhD the- sis, The Pennsylvania State University, 2017. 2

    2017

  2. [10]

    A point-based scattering model for the incoherent component of the scattered field.The Journal of the Acoustical Society of America, 141(3):EL210–EL215, 2017

    Daniel C Brown, Shawn F Johnson, and Derek R Olson. A point-based scattering model for the incoherent component of the scattered field.The Journal of the Acoustical Society of America, 141(3):EL210–EL215, 2017. 1, 2

    2017

  3. [11]

    Sediment volume search sonar development

    Daniel C Brown, Shawn Johnson, Caleb Brown- stead, Joseph Calantoni, and Gene Fabian. Sediment volume search sonar development. Technical re- port, SERDP/ESTCP, 2021.https : / / serdp - estcp . org / projects / details / 97e4a43c - bb65-432d-94ba-6aaff361571a. 5

    2021

  4. [12]

    PhD thesis, University of Can- terbury, 2003

    Hayden J Callow.Signal Processing for Synthetic Aperture Sonar Image Enhancement. PhD thesis, University of Can- terbury, 2003. 1, 2, 5

    2003

  5. [13]

    Neural implicit surface reconstruction of free- hand 3d ultrasound volume with geometric constraints.Med- ical Image Analysis, 98:103305, 2024

    Hongbo Chen, Logiraj Kumaralingam, Shuhang Zhang, Sheng Song, Fayi Zhang, Haibin Zhang, Thanh-Tu Pham, Kumaradevan Punithakumar, Edmond HM Lou, Yuyao Zhang, et al. Neural implicit surface reconstruction of free- hand 3d ultrasound volume with geometric constraints.Med- ical Im...

    2024

  6. [14]

    Airsas: Controlled dataset generation for physics-informed machine learning

    Benjamin Cowen, J Daniel Park, Thomas E Blanford, Ge- off Goehle, and Daniel C Brown. Airsas: Controlled dataset generation for physics-informed machine learning. In NeurIPS Data-Centric AI Workshop, 2021. 5

    2021

  7. [15]

    Isar-nerf: Neural radiance fields for 3-d imaging of space target from multiview isar images.IEEE Sensors Journal, 24(7):11705–11722, 2024

    Junyuan Deng, Pengfei Xie, Lei Zhang, and Yulai Cong. Isar-nerf: Neural radiance fields for 3-d imaging of space target from multiview isar images.IEEE Sensors Journal, 24(7):11705–11722, 2024. 2

    2024

  8. [16]

    Springer Science & Business Media, 2012

    Jerry Eaves and Edward Reedy.Principles of modern radar. Springer Science & Business Media, 2012. 2

    2012

  9. [17]

    Review article sar interferometry—issues, techniques, applications.Interna- tional Journal of Remote Sensing, 17(10):1803–1835, 1996

    Rudiger Gens and John L Van Genderen. Review article sar interferometry—issues, techniques, applications.Interna- tional Journal of Remote Sensing, 17(10):1803–1835, 1996. 2

    1996

  10. [18]

    Gpu acceleration for synthetic aperture sonar image reconstruction

    Isaac D Gerg, Daniel C Brown, Stephen G Wagner, Daniel Cook, Brian N O’Donnell, Thomas Benson, and Thomas C Montgomery. Gpu acceleration for synthetic aperture sonar image reconstruction. InGlobal Oceans 2020: Singapore– US Gulf Coast, pages 1–9. IEEE, 2020. 1, 2

    2020

  11. [19]

    Blanford, J

    Geoff Goehle, Benjamin Cowen, Thomas E. Blanford, J. Daniel Park, and Daniel C. Brown. Approximate extrac- tion of late-time returns via morphological component anal- ysis.The Journal of the Acoustical Society of America, 153 (5):2838, 2023. 5

    2023

  12. [20]

    Interferometric synthetic aperture sonar for high-resolution 3-d mapping of the seabed.IEE Proceedings-Radar, Sonar and Navigation, 144(2):96–103,

    HD Griffiths, TA Rafik, Z Meng, CFN Cowan, H Shafeeu, and DK Anthony. Interferometric synthetic aperture sonar for high-resolution 3-d mapping of the seabed.IEE Proceedings-Radar, Sonar and Navigation, 144(2):96–103,

  13. [21]

    Signal processing for auv based inter- ferometric synthetic aperture sonar.IEEE Oceans 2003, 5: 2438–2444, 2003

    Roy Edgar Hansen, Torstien Olsom Saebo, Kenneth Gade, and Sean Chapman. Signal processing for auv based inter- ferometric synthetic aperture sonar.IEEE Oceans 2003, 5: 2438–2444, 2003. 2

    2003

  14. [22]

    Challenges in seafloor imaging and mapping with synthetic aperture sonar.IEEE Transactions on Geoscience and Remote Sensing, 49(10): 3677–3687, 2011

    Roy Edgar Hansen, Hayden John Callow, Torstein Olsmo Sabo, and Stig Asle Vaksvik Synnes. Challenges in seafloor imaging and mapping with synthetic aperture sonar.IEEE Transactions on Geoscience and Remote Sensing, 49(10): 3677–3687, 2011. 1

    2011

  15. [23]

    Synthetic aperture sonar: A review of current status.IEEE Journal of Oceanic Engineering, 34(3):207–224, 2009

    Michael P Hayes and Peter T Gough. Synthetic aperture sonar: A review of current status.IEEE Journal of Oceanic Engineering, 34(3):207–224, 2009. 1, 2

    2009

  16. [24]

    Dart: Im- plicit doppler tomography for radar novel view synthesis

    Tianshu Huang, John Miller, Akarsh Prabhakara, Tao Jin, Tarana Laroia, Zico Kolter, and Anthony Rowe. Dart: Im- plicit doppler tomography for radar novel view synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 24118–24129, 2024. 2

    2024

  17. [25]

    Douglas M. Jennewein, Johnathan Lee, Chris Kurtz, William Dizon, Ian Shaeffer, Alan Chapman, Alejandro Chi- quete, Josh Burks, Amber Carlson, Natalie Mason, Arhat Kobawala, Thirugnanam Jagadeesan, Praful Bhargav Basani, Torey Battelle, Rebecca Belshe, Deb McCaffrey, Marisa Bra...

    2023

  18. [26]

    Air-based synthetic aperture sonar tomography using compressive sensing

    Josiah Jideani and Andrew Wilkinson. Air-based synthetic aperture sonar tomography using compressive sensing. In 2014 IEEE International Ultrasonics Symposium, pages 49–

    2014

  19. [27]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, 2015. 5

    2015

  20. [28]

    Lorensen and Harvey E

    William E. Lorensen and Harvey E. Cline. Marching cubes: A high resolution 3d surface construction algorithm. InPro- ceedings of SIGGRAPH ’87, pages 163–169, 1987. 5

    1987

  21. [29]

    Transien- tangelo: Few-viewpoint surface reconstruction using single- photon lidar

    Weihan Luo, Anagh Malik, and David B Lindell. Transien- tangelo: Few-viewpoint surface reconstruction using single- photon lidar. InProceedings of the Winter Conference on Applications of Computer Vision (WACV), pages 8712–8722,

  22. [30]

    Malik, P

    A. Malik, P. Mirdehghan, S. Nousias, K. N. Kutulakos, and D. B. Lindell. Transient neural radiance fields for lidar view synthesis and 3d reconstruction. InProc. NeurIPS, 2023. 2

    2023

  23. [31]

    Coherent and semi-coherent processing of limited-aperture circular synthetic aperture (csas) data

    Timothy M Marston, Jermaine L Kennedy, and Philip L Marston. Coherent and semi-coherent processing of limited-aperture circular synthetic aperture (csas) data. In OCEANS’11 MTS/IEEE KONA, pages 1–6. IEEE, 2011. 2

    2011

  24. [32]

    Autofocusing circular synthetic aperture sonar im- agery using phase corrections modeled as generalized cones

    Timothy M Marston, Jermaine L Kennedy, and Philip L Marston. Autofocusing circular synthetic aperture sonar im- agery using phase corrections modeled as generalized cones. The Journal of the Acoustical Society of America, 136(2): 614–622, 2014. 2

    2014

  25. [33]

    Srinivasan, Matthew Tancik, Jonathan T

    Ben Mildenhall, Pratul P. Srinivasan, Matthew Tancik, Jonathan T. Barron, Ravi Ramamoorthi, and Ren Ng. Rep- resenting scenes as neural radiance fields for view synthesis. InECCV, 2020. 2, 4

    2020

  26. [34]

    Springer, 2006

    Claus M ¨uller.Spherical harmonics. Springer, 2006. 4

    2006

  27. [35]

    Instant neural graphics primitives with a multires- olution hash encoding.ACM TOG, 41(4):1–15, 2022

    Thomas M ¨uller, Alex Evans, Christoph Schied, and Alexan- der Keller. Instant neural graphics primitives with a multires- olution hash encoding.ACM TOG, 41(4):1–15, 2022. 3

    2022

  28. [36]

    Nerf-enabled analysis-through- synthesis for isar imaging of small everyday objects with sparse and noisy uwb radar data

    Md Farhan Tasnim Oshim, Albert Reed, Suren Jayasuriya, and Tauhidur Rahman. Nerf-enabled analysis-through- synthesis for isar imaging of small everyday objects with sparse and noisy uwb radar data. In2024 IEEE/RSJ Interna- tional Conference on Intelligent Robots and Systems (I...

    2024

  29. [37]

    Alternative representations and object clas- sification of circular synthetic aperture in-air acoustic data

    J Daniel Park, Thomas E Blanford, Daniel C Brown, and Daniel Plotnick. Alternative representations and object clas- sification of circular synthetic aperture in-air acoustic data. The Journal of the Acoustical Society of America, 148(4): 2661–2661, 2020. 5

    2020

  30. [38]

    Fast nearfield to farfield conversion algorithm for circular synthetic aperture sonar.The Journal of the Acous- tical Society of America, 136(2):EL61–EL66, 2014

    Daniel S Plotnick, Philip L Marston, and Timothy M Marston. Fast nearfield to farfield conversion algorithm for circular synthetic aperture sonar.The Journal of the Acous- tical Society of America, 136(2):EL61–EL66, 2014. 2

    2014

  31. [39]

    Neural implicit surface reconstruction using imaging sonar

    Mohamad Qadri, Michael Kaess, and Ioannis Gkioulekas. Neural implicit surface reconstruction using imaging sonar. In2023 IEEE International Conference on Robotics and Au- tomation (ICRA), pages 1040–1047. IEEE, 2023. 2

    2023

  32. [40]

    Aoneus: A neural rendering framework for acoustic-optical sensor fusion

    Mohamad Qadri, Kevin Zhang, Akshay Hinduja, Michael Kaess, Adithya Pediredla, and Christopher A Metzler. Aoneus: A neural rendering framework for acoustic-optical sensor fusion. InACM SIGGRAPH 2024 Conference Papers, pages 1–12, 2024. 2

    2024

  33. [41]

    A signal-processing framework for inverse rendering

    Ravi Ramamoorthi and Pat Hanrahan. A signal-processing framework for inverse rendering. InProceedings of the 28th Annual Conference on Computer Graphics and Interactive Techniques, pages 117–128, 2001. 4

    2001

  34. [42]

    Brown, and Suren Jayasuriya

    Albert Reed, Thomas Blanford, Daniel C. Brown, and Suren Jayasuriya. Sinr: Deconvolving circular sas images using implicit neural representations.IEEE Journal of Selected Topics in Signal Processing, pages 1–16, 2022. 2

    2022

  35. [43]

    A. Reed, J. Kim, T. Blanford, A. Pediredla, D. Brown, and S. Jayasuriya. Neural volumetric reconstruction for coherent synthetic aperture sonar.ACM Trans. Graph., 42(4), 2023. 1, 2, 3, 4, 5, 6, 7, 8

    2023

  36. [44]

    Synthetic aper- ture sonar in air medium using a nonlinear sidelobe canceler

    Hiroshi Saruwatari and Mitsuo Komura. Synthetic aper- ture sonar in air medium using a nonlinear sidelobe canceler. Electronics and Communications in Japan (Part III: Funda- mental Electronic Science), 82(12):13–25, 1999. 1

    1999

  37. [45]

    Spherical harmonics.The American Math- ematical Monthly, 73(4P2):115–121, 1966

    Robert T Seeley. Spherical harmonics.The American Math- ematical Monthly, 73(4P2):115–121, 1966. 4

    1966

  38. [46]

    Non-line-of- sight imaging via neural transient fields.IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(7):2257– 2268, 2021

    Siyuan Shen, Zi Wang, Ping Liu, Zhengqing Pan, Ruiqian Li, Tian Gao, Shiying Li, and Jingyi Yu. Non-line-of- sight imaging via neural transient fields.IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(7):2257– 2268, 2021. 2

    2021

  39. [47]

    Processing video-sar data with the fast backprojection method.IEEE Transactions on Aerospace and Electronic Systems, 52(6):2838–2848, 2016

    Xiaoshen Song and Weidong Yu. Processing video-sar data with the fast backprojection method.IEEE Transactions on Aerospace and Electronic Systems, 52(6):2838–2848, 2016. 2

    2016

  40. [48]

    New York: Wiley, 1999

    Mehrdad Soumekh.Synthetic aperture radar signal process- ing. New York: Wiley, 1999. 2

    1999

  41. [49]

    Spinr: Neural volumetric reconstruction for fmcw radars.arXiv preprint arXiv:2503.23313, 2025

    Harshvardhan Takawale and Nirupam Roy. Spinr: Neural volumetric reconstruction for fmcw radars.arXiv preprint arXiv:2503.23313, 2025. 2

    2025 arXiv

  42. [50]

    Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction

    Peng Wang, Lingjie Liu, Yuan Liu, Christian Theobalt, Taku Komura, and Wenping Wang. Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction. InProceedings of the 35th International Conference on Neu- ral Information Processing Systems, Red Ho...

  43. [51]

    Bathymetric surveying with imaging sonar using neural vol- ume rendering.IEEE Robotics and Automation Letters, 9(9): 8146–8153, 2024

    Yiping Xie, Giancarlo Troni, Nils Bore, and John Folkesson. Bathymetric surveying with imaging sonar using neural vol- ume rendering.IEEE Robotics and Automation Letters, 9(9): 8146–8153, 2024. 2 SH-SAS: An Implicit Neural Representation for Complex Spherical-Harmonic Scatteri...

    2024

  44. [53]

    ellipsoid frame

    Ellipsoidal Sampling: Derivation and Ray–Ellipsoid Intersection Our measurements are bistatic time-of-flight (ToF), so each sam- ple constrains contributing scene points to a two-focus ellipsoid with foci at the TX/RX. The correct way to associate a ToF bin with spatial hypoth...

  45. [54]

    8: backprojection exhibits speckle and fragmentation, while Reed et al

    Simulations—Additional Results Qualitatively, our reconstructions are cleaner and more complete across all four synthetic scenes (Armadillo, Bunny, Happy Bud- dha, XYZ Dragon) Fig. 8: backprojection exhibits speckle and fragmentation, while Reed et al. tends to over-smooth and...

  46. [55]

    An SHlevelLin- cludes all degreesℓ= 0,

    Ablation Study: SH levels Background on spherical harmonics.Spherical harmon- ics (SH) form an orthonormal basis on the unit sphereS 2 and let us represent directional or view-dependent terms (e.g., scatter- ing/BRDF lobes) as a band-limited expansion. An SHlevelLin- cludes al...

  47. [56]

    Effect of Threshold on Visualization Figure 9 examines how sensitive each method is to the marching- cubes threshold used to extract meshes from the learned field. As the threshold increases (top to bottom), Backprojection swings from an overfilled slab-like volume to severe e...

  48. [57]

    Signal fitting and novel-view transient syn- thesis We evaluate the network’s ability to fit measured signals and then use that fit to synthesize transients at unseen poses. We define novel-view transient synthesisas generating the complex tran- sient (analytic signal—real and...

  49. [2021]

    Curran Associates Inc. 4

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.