Pith. sign in

REVIEW 3 major objections 6 minor 58 references

Hyperspectral Intrinsic Decomposition: Joint Recovery of Reflectance and Photometric Components for Non-Lambertian Scenes

T0 review · 3 major / 6 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read Hyperspectral intrinsic decomposition in non-Lambertian scenes can be reduced to estimating two spectral–spatial variables, with illuminant, shading, and specularity derived in closed form, making complete blind decomposition from one image

desk verdict A solid non-Lambertian HID framework with a genuinely useful dataset, but the quantitative leaderboard is built on synthetic data from the authors' own generator and the reflectance scale-ambiguity claim in Sec. VI-B is analytically wrong. read the letter →

arxiv 2607.25371 v1 pith:BY57RIT3 submitted 2026-07-28 cs.CV

Hao Ye , Zhan Shi , Chenglong Huang , Tao Lv , Mingjie Ji , Qiu Shen , Xun Cao This is my paper · ORCID
classification cs.CV
keywords hyperspectralintrinsicdecompositiondichromaticreflectionmodelnon-LambertianscenesspecularreflectancerecoveryilluminantestimationDichroicFormerimage
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

The paper sets out to show that non-Lambertian hyperspectral intrinsic decomposition—separating a hyperspectral image into material reflectance plus photometric effects of shading, specular reflection, and illuminant—does not require estimating four heterogeneous components. It reformulates the dichromatic reflection model so that only two spectral–spatial variables, the diffuse term and the reflectance, need to be predicted; the illuminant spectrum, shading factor, and specular coefficient then follow in closed form. On this basis it builds DichroicFormer, a dual-scale network whose global stage uses photometrically invariant gradient-ratio descriptors to preserve intrinsic boundaries and whose local stage uses specularity-guided attention to refine strong-highlight and clipped regions. Supported by a new real-world annotated dataset and a controllable synthetic generator, the paper reports that DichroicFormer outperforms existing intrinsic-decomposition baselines on every evaluated component under both standard and scale-coupled metrics, making complete blind decomposition from a single image plausible.

What carries the argument

The load-bearing identity is the two-variable inversion of the dichromatic reflection model: with P_d(u,λ)=g(u)L(λ)P_r(u,λ) and P_s=I−P_d, the four coupled components are re-expressed as two spectral–spatial targets, and closed-form averaging gives the illuminant spectrum, shading factor, and specular coefficient. Two derived photometrically invariant descriptors carry the global stage: the Spectral-Gradient Ratio (SR), which suppresses illuminant and specular terms while retaining shading-modulated reflectance differences, and the Cross Spectral-Gradient Ratio (CSR), which depends only on reflectance; they are injected as boundary priors aligned to the diffuse and reflectance subnetworks. T

What would settle it

Acquire a set of real objects with the same cross-polarization measurement protocol used to build the paper's dataset, but under illuminants and specular configurations outside the synthetic generator's sampling pools, compute pixel-wise ground truth for reflectance, shading, and specular coefficient, and check whether DichroicFormer's roughly 5–13× MSE advantages over the baselines persist; if the margins vanish, the synthetic training distribution—not the inversion math—is doing the work.

Watch

Extended reading notes

Core claim

Under the neutral-interface dichromatic reflection model, an observed hyperspectral pixel is I(u,λ)=g(u)L(λ)S(u,λ)+k(u)L(λ). The key algebraic observation is that defining the diffuse term P_d=gL S and keeping reflectance P_r=S turns the image into I=P_d+P_s, so the specular term is just I−P_d. Estimating the two spectral–spatial targets P_d and P_r therefore determines the full factorization: the illuminant spectrum is recovered by spatially averaging P_d/P_r plus P_s with peak normalization, and the shading factor and specular coefficient are recovered by spectral averaging. This reduction removes the dimensional mismatch that forced earlier methods to estimate components separately, and t

Load-bearing premise

The quantitative results rest on the assumption that the paper's synthetic training-and-test renderer—a simple linear mixture of diffuse and specular terms followed by clipping—faithfully captures real imaging, since real-capture evaluation is only qualitative.

Editorial extensions

If this is right

  • A single hyperspectral image, without depth, multi-view, or controlled illumination hardware, can in principle yield a complete dichromatic decomposition: reflectance, shading, specular coefficient, and illuminant spectrum.
  • Because the illuminant, shading, and specular coefficient are derived rather than predicted, their relative scales are automatically consistent with the image formation model, which is why the paper introduces shared-scale rs-MSE metrics.
  • Decoupling the specular output from the analytic I−P_d path allows the method to handle saturated and clipped specular highlights without letting clipping corrupt the diffuse estimate.
  • The synthetic generator's independent control over illuminant spectra and diffuse–specular balance enables training and testing under conditions real captures cannot exhaust, while the new real-world annotations provide a reference for non-Lambertian evaluation.
  • The two-variable reformulation is not network-specific: any estimator of P_d and P_r inherits the closed-form recovery of the photometric components.

Reading between the lines

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

  • If the inversion reduction is as domain-simplifying as it appears, the same P_d/P_r parametrization could be applied to RGB or multispectral intrinsic decomposition, since the closed-form steps are wavelength-agnostic; this is an extension the paper does not claim.
  • The large gap between si-MSE and rs-MSE for the baselines suggests that per-component scale alignment has been systematically hiding relative-scale errors; adopting shared-scale metrics more widely in intrinsic imaging would likely re-rank existing methods—an editorial inference, not a paper claim.
  • The SR/CSR descriptors assume high spectral sampling density so the illuminant is locally flat; a natural stress test is whether performance degrades gracefully on coarse-band or strongly peaked-illuminant sensors, which the paper leaves to future work.
  • The generator's independent control over diffuse and specular scaling could be used to map robustness curves, such as error versus specular fraction or clipping severity, giving practitioners a principled way to choose acquisition exposure settings.
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

3 major / 6 minor

Summary. The paper proposes DichroicFormer, a single-image hyperspectral intrinsic decomposition method for non-Lambertian scenes under the dichromatic reflection model (DRM). The core idea is to reformulate the recovery of four coupled components (reflectance, shading factor, specular coefficient, illuminant spectrum) as estimation of two spectral–spatial variables: the diffuse term P_d and the reflectance P_r. Illuminant L, shading g, and specular coefficient k are then derived in closed form. The architecture is a dual-scale network with a global-stage invariant-driven module using spectral-gradient ratios and a local-stage specularity-guided attention module. The paper also introduces the CITE real-world dataset with component-level annotations and the PISG synthetic generator. Quantitative evaluation on PISG-generated test pairs is claimed to show state-of-the-art performance under both scale-invariant and scale-coupled metrics, with qualitative results on CITE and KAUST real HSIs.

Significance. If the claims hold, the paper makes several valuable contributions: a useful algebraic reduction of DRM inversion, a new public dataset (CITE), a controllable physically motivated generator (PISG), and a carefully ablated architecture. The closed-form derivations for L, g, and k are parameter-free and transparent, and the dual-scale design with photometrically invariant descriptors is well motivated. The ablations are internally consistent and isolate contributions of the proposed modules. However, the paper's central quantitative claim is weakened by two issues: the absolute reflectance metric is gauge-dependent because of a genuine scale ambiguity in the DRM, and all quantitative results are obtained on PISG-synthesized test images rather than on real captures. These issues do not necessarily invalidate the method, but they require correction or additional evidence before the claimed state-of-the-art performance can be accepted.

major comments (3)
  1. [Sec. VI-B, Eq. (1)] The statement "For reflectance, which admits no scale ambiguity under the joint constraint, we report absolute MSE and SSIM" is analytically incorrect under the DRM. In Eq. (1), the transformation g(u) -> alpha g(u), S(u,lambda) -> S(u,lambda)/alpha leaves I(u,lambda) unchanged for any alpha>0, with L and k untouched. This is an exact symmetry of the image-formation model. The closed-form inversion in Eqs. (4)-(5) preserves this ambiguity rather than removing it. Consequently, absolute MSE for reflectance depends on the gauge in which each method happens to output S, and the large gap in Table I (0.007 vs 0.189) may substantially reflect gauge alignment of the learned estimator to the PISG training convention rather than physical decomposition fidelity. The authors should either adopt a canonical gauge (e.g., normalize each predicted reflectance by a fixed scalar) and apply it to all met
  2. [Secs. V-B, VI-A, VI-C; Eq. (16)] All quantitative results in Tables I-III are produced on PISG-generated test pairs: "the PISG generates 240 test pairs from the held-out objects" (Sec. VI-A), using the same DRM-plus-clipping synthesis model (Eq. (16)) as the training distribution. Real CITE captures are evaluated only qualitatively (Sec. VI-D.1, VI-D.2), and the KAUST experiment is also qualitative. Thus the central SOTA claim is not yet supported by evidence on real sensor data, where noise, misalignment, polarizer imperfections, and deviations from the DRM may substantially change performance. I recommend adding a quantitative evaluation on real CITE images using the component annotations obtained from Eq. (15), or explicitly redrawing the quantitative claims to the synthetic distribution. At minimum, the paper should report how the PISG test distribution relates to the real CITE captures (e.g., how much clipping, noi
  3. [Sec. IV-C; Eq. (3)] The paper replaces the analytic relation P_s = I - P_d with a learned specular prediction P_L_s from SGAM, and the two are decoupled: "the analytic pathway ties the final specular output directly to the diffuse estimate" and the specular branch is independently supervised and detached. This means the final outputs do not have to satisfy P_L_d + P_L_s = I, so the closed-form derivations of L, g, and k in Eqs. (4)-(5), which assume a consistent P_s, may not be anchored to the observed image. The recomposition panels in Fig. 8 are visual only. I recommend reporting a quantitative recomposition error, e.g., ||I - (P_L_d + P_L_s)||, on the test set and on real captures, and stating explicitly whether the derived photometric components are required to satisfy the DRM exactly or only approximately.
minor comments (6)
  1. [Sec. IV-A] The notation "P^G_d = I + R^G_s" appears inconsistent with the standard relation P_d = I - P_s. If R^G_s is a negative specular residual, this should be stated; otherwise the sign convention is confusing.
  2. [Sec. V-A, Eq. (15)] The expression "R = I_perp ⊙ R_coat / I_coat" mixes element-wise multiplication and division in an ambiguous way. Please use explicit element-wise division or bracket notation.
  3. [Sec. IV-B, Eqs. (9)-(12)] The approximation rho_1 ≈ rho_2 ≈ 1 is justified by a small spectral sampling interval. Since the paper subsamples to 36 bands over 440-700 nm, the effective interval is roughly 7-8 nm. Please state whether the SR/CSR approximation was validated at this subsampled resolution, or whether the descriptors are computed on the full 80-band data before subsampling.
  4. [Sec. VI-A] The training illuminant pool is described as containing 30 spectra, while the test set uses five illuminant spectra. Please clarify whether these five are seen during training or are held-out, and how they were selected.
  5. [Sec. VI-C] The sentence "DichroicFormer achieves the best performance on every evaluated component under both metric families" is too strong given the gauge-dependence of the absolute MSE metric. Please qualify it in view of the scale ambiguity discussed above.
  6. [General] Several formal analyses are deferred to the Supplementary Material, including the formal argument about scale relationships. Since the main text makes strong claims about scale ambiguity, the key statement should at least be summarized in the main text or the supplementary material should be provided with the submission.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation; the closed-form inversion formulas are algebraic identities from the DRM

full rationale

The central inversion paradigm (Eqs. 2-5) is a reparameterization, not a circular prediction: P_d and P_r are defined as gLS and S, and L, g, k are then obtained by algebraic manipulation of P_d, P_r, P_s, with P_s = I - P_d. No target quantity is fed back into the definitions; the network's actual job is to estimate the two spectral-spatial target variables, after which the remaining components are computed in closed form. The SR/CSR descriptors are derived from the DRM and used only as conditioning priors, not as supervised targets, so they do not smuggle in the output. The quantitative test uses PISG-generated pairs from the same generative equation (Eq. 16) as the training distribution, which is a data-distribution limitation rather than a circular step; held-out objects and unseen illuminant/config combinations still make this a legitimate within-distribution benchmark, and the external KAUST results are qualitative only. The claim that reflectance has no scale ambiguity under the joint constraint is analytically questionable, since g -> αg and S -> S/α leave Eq. 1 invariant, making the absolute-MSE metric gauge-sensitive; however, this is a metric-validity concern, not a reduction of the derivation to its inputs. Self-citations in the related work are not load-bearing for the proposed derivation.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

No new physical entities, forces, or particles are postulated. The free parameters are data-generation and preprocessing choices. The key unstated burden is the PISG distributional assumption, since both training and the quantitative test set come from that generator.

free parameters (3)
  • PISG diffuse scaling range a = sampled uniformly from [0.65, 0.85]
    Hand-chosen range for data augmentation in the Physically-faithful Intrinsic Set Generator; affects training distribution and, indirectly, all quantitative results.
  • PISG specular scaling range b = sampled uniformly from [0.5, 1.1]
    Hand-chosen range for specular intensity in synthesized training/test pairs; affects the balance of diffuse/specular configurations.
  • Spectral band selection = 36 bands from 80 (440-700 nm), short-wavelength end below 450 nm discarded
    Data preprocessing choice stated in Sec. VI-A; affects all results and the descriptors that depend on spectral sampling.
assumptions (5)
  • domain assumption Dichromatic reflection model with neutral interface reflection and a single illuminant spectral distribution (Eq. 1).
    Stated at the start of Sec. III; all subsequent derivations assume this image formation model.
  • domain assumption Illuminant spectrum is approximately constant over adjacent spectral channels (rho1 ≈ rho2 ≈ 1), making the residual epsilon in Eq. (11) negligible.
    Load-bearing for the SR/CSR invariant descriptors in Sec. IV-B; fails in regions with strong specularity or coarse spectral sampling, as the authors acknowledge.
  • domain assumption Cross-polarization capture fully suppresses specular reflection and the lacquer coating provides a uniform reference reflectance, so Eq. (15) yields ground-truth R, g, k.
    Underlies all CITE annotations; requires that coating and polarization do not alter geometry, illumination, or diffuse reflectance.
  • domain assumption PISG-synthesized images (Eqs. 16-17) are representative of real non-Lambertian captures.
    Quantitative evaluation and training both rely on PISG; if the synthetic renderer omits real sensor noise, BRDF departures, or spatially varying illumination, the reported margins may not transfer.
  • domain assumption The learned network can infer the non-unique decomposition from a single image given supervised training.
    The DRM inversion is underdetermined; the method relies on learned priors rather than a closed-form unique solution.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hyperspectral Intrinsic Decomposition: Joint Recovery of Reflectance and Photometric Components for Non-Lambertian Scenes." pith.science (2026). https://pith.science/paper/BY57RIT3

@misc{pith2026260725371,
  author       = {Pith},
  title        = {Pith review of: Hyperspectral Intrinsic Decomposition: Joint Recovery of Reflectance and Photometric Components for Non-Lambertian Scenes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BY57RIT3}},
  note         = {Machine review of arXiv:2607.25371}
}
read the original abstract

Hyperspectral intrinsic decomposition (HID) aims to disentangle material-related spectral properties and photometric effects in hyperspectral images (HSIs), which is essential for understanding real-world imaging processes and benefits a variety of downstream applications. Most existing HID studies have been developed under Lambertian or near-Lambertian assumptions. The few prior non-Lambertian efforts rely on simplified specular assumptions insufficient to handle diverse real-world specularity, and typically require auxiliary inputs or recover only a subset of the coupled reflectance and photometric components, hindering complete and blind decomposition. In this paper, we revisit the dichromatic reflection model (DRM) and develop a unified inversion paradigm that reformulates the recovery of four coupled reflectance and photometric components as the estimation of two spectral--spatial target variables. Building on this reformulation, we propose a dual-scale decomposition scheme to handle non-Lambertian effects with distinct spatial characteristics. At the global scale, photometrically invariant descriptors serve as edge priors for high-fidelity intrinsic boundary preservation; at the local scale, specularity-guided attention directs refinement with emphasis on specularity-dominated regions, including those affected by clipping distortion. To facilitate future research, we establish CITE, the first public real-world HID dataset for non-Lambertian objects, and develop a Physically-faithful Intrinsic Set Generator (PISG) for controllable data synthesis. Extensive ablation studies and experiments on the CITE and additional HSIs demonstrate the effectiveness of our method and its robustness across diverse scenes.

Figures

Figures reproduced from arXiv: 2607.25371 by the authors.

Figure 1
Figure 1. Hyperspectral intrinsic decomposition on non-Lambertian surfaces. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed dual-scale decomposition framework. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. HSI and the corresponding SR and CSR visualizations for two representative scenes. Arrows indicate selected local details for comparison. The red box highlights a region where soft specularity attenuates inter-material spectral differences in the original HSI; the zoomed insets show that both SR and CSR reveal these contrasts more clearly, as confirmed by comparison with the ground-truth reflectance (green border). … view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Overview of the CITE data acquisition pipeline and representative annotations. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison of reflectance recovery on two CITE test objects. Each object is shown with three complementary views: a three-dimensional [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Shading factor estimates on two CITE test objects. Results are shown as grayscale visualizations. Zoomed insets marked in [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Specular coefficient estimates on two CITE test objects. This comparison includes both intrinsic decomposition baselines and task-specific specular [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Component-scale analysis on CITE. For each of the two test objects, the statistical panels summarize the distributions of the decomposed components, [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: Illuminant estimation on CITE. Normalized spectra estimated by [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: Qualitative intrinsic decomposition results on the KAUST dataset for two non-Lambertian scenes: (a) tomatoes with fine-grained pigment variations [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 1 linked inside Pith

  1. [1]

    Aperture diffraction for compact snapshot spectral imaging,

    T. Lv, H. Ye, Q. Yuan, Z. Shi, Y . Wang, S. Wang, and X. Cao, “Aperture diffraction for compact snapshot spectral imaging,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 10 574–10 584

    2023

  2. [2]

    Compact self-adaptive coding for spectral compressive sensing,

    Z. Shi, H. Ye, T. Lv, Y . Wang, and X. Cao, “Compact self-adaptive coding for spectral compressive sensing,” in2023 IEEE International Conference on Computational Photography (ICCP). IEEE, 2023, pp. 1–12

    2023

  3. [3]

    Evaluation of prospect inversion approaches using multi-angular spectral reflectance factor of leaves,

    C. Yao, Z. Sun, and S. Lu, “Evaluation of prospect inversion approaches using multi-angular spectral reflectance factor of leaves,”IEEE Trans- actions on Geoscience and Remote Sensing, 2024

    2024

  4. [4]

    Decomposing multispectral face images into diffuse and specular shading and biophysical parameters,

    S. Alotaibi and W. A. Smith, “Decomposing multispectral face images into diffuse and specular shading and biophysical parameters,” in2019 IEEE International Conference on Image Processing (ICIP). IEEE, 2019, pp. 3138–3142

    2019

  5. [5]

    Multispectral illumination estimation using deep unrolling network,

    Y . Li, Q. Fu, and W. Heidrich, “Multispectral illumination estimation using deep unrolling network,” inProceedings of the IEEE/CVF inter- national conference on computer vision, 2021, pp. 2672–2681

    2021

  6. [6]

    Multispectral photometric stereo for acquiring high-fidelity surface normals,

    G. Nam and M. H. Kim, “Multispectral photometric stereo for acquiring high-fidelity surface normals,”IEEE computer graphics and applica- tions, vol. 34, no. 6, pp. 57–68, 2014

    2014

  7. [7]

    Intrinsic decomposition from a single spectral image,

    X. Chen, W. Zhu, Y . Zhao, Y . Yu, Y . Zhou, T. Yue, S. Du, and X. Cao, “Intrinsic decomposition from a single spectral image,”Applied optics, vol. 56, no. 20, pp. 5676–5684, 2017

    2017

  8. [8]

    Multispectral image intrinsic decomposition via subspace constraint,

    Q. Huang, W. Zhu, Y . Zhao, L. Chen, Y . Wang, T. Yue, and X. Cao, “Multispectral image intrinsic decomposition via subspace constraint,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 6430–6439

    2018

Show all 58 references
  1. [9]

    Optimal solution of the dichromatic model for multispectral photometric invariance,

    C. P. Huynh and A. Robles-Kelly, “Optimal solution of the dichromatic model for multispectral photometric invariance,” inStructural, Syntactic, and Statistical Pattern Recognition: Joint IAPR International Workshop, SSPR & SPR 2008, Orlando, USA, December 4-6, 2008. Proceeding...

    2008

  2. [10]

    A solution of the dichromatic model for multispectral photometric invariance,

    ——, “A solution of the dichromatic model for multispectral photometric invariance,”International Journal of Computer Vision, vol. 90, no. 1, pp. 1–27, 2010

    2010

  3. [11]

    Efficient estimation of reflectance parameters from imaging spectroscopy,

    L. Gu, A. A. Robles-Kelly, and J. Zhou, “Efficient estimation of reflectance parameters from imaging spectroscopy,”IEEE Transactions on Image Processing, vol. 22, no. 9, pp. 3648–3663, 2013

    2013

  4. [12]

    Shape, illumination, and reflectance from shading,

    J. T. Barron and J. Malik, “Shape, illumination, and reflectance from shading,”IEEE transactions on pattern analysis and machine intelli- gence, vol. 37, no. 8, pp. 1670–1687, 2014

    2014

  5. [13]

    Intrinsic rgb and multispectral images recovery by independent quadratic programming,

    A. Krebs, Y . Benezeth, and F. Marzani, “Intrinsic rgb and multispectral images recovery by independent quadratic programming,”PeerJ Com- puter Science, vol. 6, p. e256, 2020

    2020

  6. [14]

    Using color to separate reflection components,

    S. A. Shafer, “Using color to separate reflection components,”Color Research & Application, vol. 10, no. 4, pp. 210–218, 1985

    1985

  7. [15]

    A benchmark dataset and evaluation for non-lambertian and uncalibrated photometric stereo,

    B. Shi, Z. Wu, Z. Mo, D. Duan, S.-K. Yeung, and P. Tan, “A benchmark dataset and evaluation for non-lambertian and uncalibrated photometric stereo,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 3707–3716

    2016

  8. [16]

    Towards high-resolution specular highlight detection,

    G. Fu, Q. Zhang, L. Zhu, Q. Lin, Y . Wang, S. Fan, and C. Xiao, “Towards high-resolution specular highlight detection,”International Journal of Computer Vision, vol. 132, no. 1, pp. 95–117, 2024

    2024

  9. [17]

    Estimating reflectance parameters of an object from saturated spectral images,

    S. Li, Y . Manabe, and K. Chihara, “Estimating reflectance parameters of an object from saturated spectral images,” inInternational Conference on Computer Graphics, Imaging and Visualisation (CGIV’06). IEEE, 2006, pp. 335–340

    2006

  10. [18]

    Saturated signals in spectroscopic imaging: why and how should we deal with this regularly observed phenomenon?

    A. Nardecchia, V . Motto-Ros, and L. Duponchel, “Saturated signals in spectroscopic imaging: why and how should we deal with this regularly observed phenomenon?”Analytica Chimica Acta, vol. 1157, p. 338389, 2021

    2021

  11. [19]

    Recovering intrinsic scene characteristics,

    H. Barrow, J. Tenenbaum, A. Hanson, and E. Riseman, “Recovering intrinsic scene characteristics,”Comput. vis. syst, vol. 2, no. 3-26, p. 2, 1978

    1978

  12. [20]

    Lightness and retinex theory,

    E. H. Land and J. J. McCann, “Lightness and retinex theory,”Josa, vol. 61, no. 1, pp. 1–11, 1971

    1971

  13. [21]

    Deriving intrinsic images from image sequences,

    Y . Weiss, “Deriving intrinsic images from image sequences,” inPro- ceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001, vol. 2. IEEE, 2001, pp. 68–75

    2001

  14. [22]

    Re- covering intrinsic images with a global sparsity prior on reflectance,

    C. Rother, M. Kiefel, L. Zhang, B. Sch ¨olkopf, and P. Gehler, “Re- covering intrinsic images with a global sparsity prior on reflectance,” Advances in neural information processing systems, vol. 24, 2011

    2011

  15. [23]

    Towards high-quality intrinsic images in the wild,

    G. Fu, Q. Zhang, and C. Xiao, “Towards high-quality intrinsic images in the wild,” in2019 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2019, pp. 175–180

    2019

  16. [24]

    A closed-form solution to retinex with nonlocal texture constraints,

    Q. Zhao, P. Tan, Q. Dai, L. Shen, E. Wu, and S. Lin, “A closed-form solution to retinex with nonlocal texture constraints,”IEEE transactions on pattern analysis and machine intelligence, vol. 34, no. 7, pp. 1437– 1444, 2012

    2012

  17. [25]

    Intrinsic image decomposition using a sparse representation of reflectance,

    L. Shen, C. Yeo, and B.-S. Hua, “Intrinsic image decomposition using a sparse representation of reflectance,”IEEE transactions on pattern analysis and machine intelligence, vol. 35, no. 12, pp. 2904–2915, 2013

    2013

  18. [26]

    Intrinsic image decomposition with non- local texture cues,

    L. Shen, P. Tan, and S. Lin, “Intrinsic image decomposition with non- local texture cues,” in2008 IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2008, pp. 1–7

    2008

  19. [27]

    Single image layer separation using relative smoothness,

    Y . Li and M. S. Brown, “Single image layer separation using relative smoothness,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2014, pp. 2752–2759

    2014

  20. [28]

    An l 1 image transform for edge-preserving smoothing and scene-level intrinsic decomposition,

    S. Bi, X. Han, and Y . Yu, “An l 1 image transform for edge-preserving smoothing and scene-level intrinsic decomposition,”ACM Transactions on Graphics (TOG), vol. 34, no. 4, pp. 1–12, 2015

    2015

  21. [29]

    Intrinsic image decom- position with step and drift shading separation,

    B. Sheng, P. Li, Y . Jin, P. Tan, and T.-Y . Lee, “Intrinsic image decom- position with step and drift shading separation,”IEEE Transactions on Visualization and Computer Graphics, vol. 26, no. 2, pp. 1332–1346, 2018

    2018

  22. [30]

    Direct intrinsics: Learning albedo- shading decomposition by convolutional regression,

    T. Narihira, M. Maire, and S. X. Yu, “Direct intrinsics: Learning albedo- shading decomposition by convolutional regression,” inProceedings of the IEEE international conference on computer vision, 2015, pp. 2992– 2992

    2015

  23. [31]

    Enhancing intrinsic image decomposition with transformer and laplacian pyramid network

    J. Liu, Y . Ma, X. Meng, S. Zhang, Z. Liu, and Y . Song, “Enhancing intrinsic image decomposition with transformer and laplacian pyramid network.”Traitement du Signal, vol. 41, no. 1, 2024

    2024

  24. [32]

    Intrinsic image transformation via scale space decomposition,

    L. Cheng, C. Zhang, and Z. Liao, “Intrinsic image transformation via scale space decomposition,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 656–665

    2018

  25. [33]

    Crefnet: Learning consis- tent reflectance estimation with a decoder-sharing transformer,

    J. Luo, N. Zhao, W. Li, and C. Richardt, “Crefnet: Learning consis- tent reflectance estimation with a decoder-sharing transformer,”IEEE Transactions on Visualization and Computer Graphics, vol. 30, no. 9, pp. 6407–6420, 2023

    2023

  26. [34]

    Self-supervised intrinsic image decomposition,

    M. Janner, J. Wu, T. D. Kulkarni, I. Yildirim, and J. Tenenbaum, “Self-supervised intrinsic image decomposition,”Advances in neural information processing systems, vol. 30, 2017

    2017

  27. [35]

    Unsupervised learning for intrinsic image decomposition from a single image,

    Y . Liu, Y . Li, S. You, and F. Lu, “Unsupervised learning for intrinsic image decomposition from a single image,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 3248–3257

    2020

  28. [36]

    Eyeir: Single eye image inverse rendering in the wild,

    S. Liang, H. Wang, and F. Lu, “Eyeir: Single eye image inverse rendering in the wild,” inACM SIGGRAPH 2024 Conference Papers, 2024, pp. 1–11

    2024

  29. [37]

    Quadratic objective functions for dichromatic model parameters estimation,

    A. Krebs, Y . Benezeth, and F. Marzani, “Quadratic objective functions for dichromatic model parameters estimation,” in2017 International Conference on Digital Image Computing: Techniques and Applications (DICTA). IEEE, 2017, pp. 1–8

    2017

  30. [38]

    An optimisation approach to the recovery of reflection parameters from a single hyperspectral image,

    S. Rahman and A. Robles-Kelly, “An optimisation approach to the recovery of reflection parameters from a single hyperspectral image,” Computer vision and image understanding, vol. 117, no. 12, pp. 1672– 1688, 2013

    2013

  31. [39]

    Estimating reflectance parameters, light direction, and shape from a single multispectral image,

    ——, “Estimating reflectance parameters, light direction, and shape from a single multispectral image,”IEEE Transactions on Computational Imaging, vol. 3, no. 4, pp. 837–852, 2017. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 18

    2017

  32. [40]

    Multi- modal non-isotropic light source modelling for reflectance estimation in hyperspectral imaging,

    J. Mehami, R. Falque, T. Vidal-Calleja, and A. Alempijevic, “Multi- modal non-isotropic light source modelling for reflectance estimation in hyperspectral imaging,”IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 10 336–10 343, 2022

    2022

  33. [41]

    Spectral dichromatic parameter recovery from two views via total variation hyper-priors,

    F. Bergamasco, A. Torsello, and A. Robles-Kelly, “Spectral dichromatic parameter recovery from two views via total variation hyper-priors,” inComputer Vision–ACCV 2016 Workshops: ACCV 2016 International Workshops, Taipei, Taiwan, November 20-24, 2016, Revised Selected Papers, ...

    2016

  34. [42]

    Spectralgaussians: Semantic, spectral 3d gaussian splatting for multi-spectral scene representation, visualization and analysis,

    S. N. Sinha, H. Graf, and M. Weinmann, “Spectralgaussians: Semantic, spectral 3d gaussian splatting for multi-spectral scene representation, visualization and analysis,”arXiv preprint arXiv:2408.06975, 2024

    2024 arXiv

  35. [43]

    Multispectral photometric stereo for spatially-varying spectral reflectances: A well posed problem?

    H. Guo, F. Okura, B. Shi, T. Funatomi, Y . Mukaigawa, and Y . Mat- sushita, “Multispectral photometric stereo for spatially-varying spectral reflectances: A well posed problem?” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 963– 971

    2021

  36. [44]

    Multispectral photometric stereo for spatially-varying spectral reflectances,

    ——, “Multispectral photometric stereo for spatially-varying spectral reflectances,”International Journal of Computer Vision, vol. 130, no. 9, pp. 2166–2183, 2022

    2022

  37. [45]

    Non-lambertian multi- spectral photometric stereo via spectral reflectance decomposition

    J. Lv, H. Guo, G. Chen, J. Liang, and B. Shi, “Non-lambertian multi- spectral photometric stereo via spectral reflectance decomposition.” in IJCAI, 2023, pp. 1249–1257

    2023

  38. [46]

    Shape and re- flectance reconstruction using concentric multi-spectral light field,

    M. Zhou, Y . Ding, Y . Ji, S. S. Young, J. Yu, and J. Ye, “Shape and re- flectance reconstruction using concentric multi-spectral light field,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 42, no. 7, pp. 1594–1605, 2020

    2020

  39. [47]

    Intrinsic image decomposition as two independent deconvolution problems,

    A. Krebs, Y . Benezeth, and F. Marzani, “Intrinsic image decomposition as two independent deconvolution problems,”Signal Processing: Image Communication, vol. 86, p. 115872, 2020

    2020

  40. [48]

    Mst++: Multi-stage spectral-wise transformer for efficient spectral reconstruction,

    Y . Cai, J. Lin, Z. Lin, H. Wang, Y . Zhang, H. Pfister, R. Timofte, and L. Van Gool, “Mst++: Multi-stage spectral-wise transformer for efficient spectral reconstruction,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 745–755

    2022

  41. [49]

    Restormer: Efficient transformer for high-resolution image restoration,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, and M.-H. Yang, “Restormer: Efficient transformer for high-resolution image restoration,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 5728–5739

    2022

  42. [50]

    Illumination intensity, object geometry and highlights invariance in multispectral imaging,

    R. Montoliu, F. Pla, and A. C. Klaren, “Illumination intensity, object geometry and highlights invariance in multispectral imaging,” inIberian Conference on Pattern Recognition and Image Analysis. Springer, 2005, pp. 36–43

    2005

  43. [51]

    Spectral invariant repre- sentation for spectral reflectance image,

    A. Ibrahim, S. Tominaga, and T. Horiuchi, “Spectral invariant repre- sentation for spectral reflectance image,” in2010 20th International Conference on Pattern Recognition. IEEE, 2010, pp. 2776–2779

    2010

  44. [52]

    Ground truth dataset and baseline evaluations for intrinsic image algorithms,

    R. Grosse, M. K. Johnson, E. H. Adelson, and W. T. Freeman, “Ground truth dataset and baseline evaluations for intrinsic image algorithms,” in 2009 IEEE 12th International Conference on Computer Vision. IEEE, 2009, pp. 2335–2342

    2009

  45. [53]

    Illuminant estimation in multispectral imaging,

    H. A. Khan, J.-B. Thomas, J. Y . Hardeberg, and O. Laligant, “Illuminant estimation in multispectral imaging,”Journal of the Optical Society of America A, vol. 34, no. 7, pp. 1085–1098, 2017

    2017

  46. [54]

    Illumination separation of non-lambertian scenes from a single hyperspectral image,

    T. Su, Y . Zhou, Y . Yu, X. Cao, and S. Du, “Illumination separation of non-lambertian scenes from a single hyperspectral image,”Optics express, vol. 26, no. 20, pp. 26 167–26 178, 2018

    2018

  47. [55]

    Shades of gray and colour constancy,

    G. D. Finlayson and E. Trezzi, “Shades of gray and colour constancy,” inColor and imaging conference, vol. 12. Society of Imaging Science and Technology, 2004, pp. 37–41

    2004

  48. [56]

    Specularity removal from imaging spec- troscopy data via entropy minimisation,

    L. Gu and A. Robles-Kelly, “Specularity removal from imaging spec- troscopy data via entropy minimisation,” in2011 International Confer- ence on Digital Image Computing: Techniques and Applications. IEEE, 2011, pp. 59–65

    2011

  49. [57]

    Separating reflection components of textured surfaces using a single image,

    R. T. Tan and K. Ikeuchi, “Separating reflection components of textured surfaces using a single image,”IEEE transactions on pattern analysis and machine intelligence, vol. 27, no. 2, pp. 178–193, 2005

    2005

  50. [58]

    Cbam: Convolutional block attention module,

    S. Woo, J. Park, J.-Y . Lee, and I. S. Kweon, “Cbam: Convolutional block attention module,” inProceedings of the European conference on computer vision (ECCV), 2018, pp. 3–19. Hao Yereceived the B.E. degree from the School of Computer Science and Engineering, Northeastern Univ...

    2018

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.