Pith. sign in

REVIEW 2 major objections 5 minor 58 references

Improved Background Estimation for Gas Plume Identification in Hyperspectral Images

T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Localized background estimation materially improves gas plume identification in longwave infrared hyperspectral images: on 640 simulated plumes, K-Nearest Segments raises median neural-network confidence in the true gas from 38.2% to…

desk verdict Useful benchmark and plausible new method, but headline confidence gains are oracle-tuned, so the practical claim is unproven. read the letter →

arxiv 2411.15378 v1 pith:KY2HXRPT submitted 2024-11-22 eess.IV cs.CVcs.LG

Scout Jarman , Zigfried Hampel-Arias , Adra Carr , Kevin R. Moon This is my paper · ORCID
classification eess.IVcs.CVcs.LG
keywords LWIRhyperspectralimaginggasplumeidentificationbackgroundestimationneuralnetworkconfidenceK-nearestsegmentsprincipalcomponentanalysiswhiteningtransformremotesensing
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 argues that the standard practice of subtracting one global average background before identifying gas plumes in LWIR hyperspectral images is a bottleneck, and that substituting local background estimates improves both background radiance accuracy and neural-network identification confidence. On 640 simulated plumes, principal component analysis gives the best background radiance estimates, with median mean squared error about 18,000 times lower than global estimation. The proposed K-Nearest Segments algorithm, which averages spectrally similar non-plume image segments found by watershed segmentation, gives the largest identification-confidence gain, raising median confidence in the true gas from 38.2% to 91.4%. The authors conclude that background estimation should be treated as a separate design choice for identification, with KNN as a robust out-of-the-box option.

What carries the argument

The load-bearing object is the background radiance estimate $L_{\text{off}}$ in the additive model. The paper's mechanism is to estimate $L_{\text{off}}$ locally, per pixel, per cluster, per principal-component subspace, per annulus, or per image segment, then whiten with the global covariance $\Sigma_g$ via $\tilde{L} = \Sigma_g^{-1/2}(L - \mu_g)$, average the whitened ROI pixels, and feed the result to a 1-D convolutional squeeze-excite neural network with a softmax output over about 10,500 materials. KNS's distinctive machinery is watershed segmentation of the image into homogeneous regions, hierarchical linkage functions (single, complete, average) to rank non-plume segments by spectral similarity to each plume segment, and an optional background-target separation step that jointly estimates the background and the gas signal by constrained optimization.

What would settle it

Rerun the 640-plume comparison reporting top-1 classification accuracy and calibration instead of softmax confidence; if KNS does not also raise top-1 accuracy over Global, or if C2H2 and N2O accuracy stays near chance despite confidence rising to 16.3% and 42.9%, the identification claim would be falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the background radiance estimate subtracted before gas identification should be local rather than global. Using the additive plume model $L = L_{\text{off}} + \psi s$, the authors argue that the standard global mean $\mu_g$ is a poor proxy for the actual background under a plume that crosses multiple surface materials, and that replacing it with per-pixel or per-segment estimates improves both the fidelity of the recovered gas signature and the confidence of a neural-network identifier. Across 640 simulated plumes spanning eight gases and eight signal-strength levels, PCA background projection produced the most accurate background radiance, with median mean squared error 18,855 times lower than Global, followed by KNN and KMeans. The proposed K-Nearest Segments method, which averages spectrally similar non-plume watershed segments, produced the highest median identification confidence, 91.4%, versus 38.2% for Global, a 53.2 percentage-point gain, and the largest gains for the hardest gases, C2H2 and N2O. The authors conclude that background estimation is a distinct design choice for identification and recommend KNN as a low-sensitivity default.

Load-bearing premise

The evaluation uses the neural network's softmax confidence in the true gas as the definition of successful identification, and the network was trained and tested on signatures generated from the same simulation model, so the 53.2-point confidence gain may not transfer to accuracy on real plumes.

Editorial extensions

If this is right

  • Operational gas identification pipelines can replace global background subtraction with KNS or Annulus to raise neural-network confidence, especially for gases like C2H2 and N2O that the network finds hard.
  • For estimating the radiance underneath a plume, PCA and KNN are the strongest choices, with KNN offering low hyperparameter sensitivity and a consistent default setting.
  • Hyperparameter defaults matter: the grid search shows small k values are generally preferred for identification, and KNS benefits from skipping background-target separation when the goal is confidence.
  • Because every method improved confidence over Global in roughly 75% of plumes, global estimation is a conservative baseline rather than a ceiling for identification performance.

Reading between the lines

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

  • If softmax confidence tracks true identification accuracy, the 53.2-point confidence gain implies KNS should also reduce misidentification rates in operational settings; the paper does not itself report accuracy, so this is an extrapolation.
  • The inverse ranking, PCA best at background radiance MSE but KNS best at confidence, suggests that minimizing background error alone is not the right objective for identification; a testable extension is to train the network with PCA-estimated backgrounds or to fuse PCA and KNS estimates.
  • The same comparison could be extended to real plumes with ground-truth gas releases; the simulation setup already includes realistic atmospheric and emissivity terms, so the main missing evidence is field data.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The paper compares five non-global background estimation methods (KMeans, PCA, KNN, Annulus, and the proposed KNS) against global background estimation for LWIR hyperspectral gas plume identification. Using 640 simulated plumes with known ground-truth background radiance, the authors report that PCA yields the lowest background MSE (median 18,855 times lower than Global) and that the proposed KNS raises the median neural-network confidence in the true gas from 38.2% to 91.4%. The methods are evaluated both by MSE to the true background and by the softmax confidence of a neural network identifier. The paper also analyzes per-plume optimal hyperparameters and sensitivity.

Significance. If the headline numbers were reliable, the paper would provide a useful practical comparison of background estimators for gas plume identification and introduce two promising methods. The simulation setup with 640 plumes, eight gases, and multiple signal strengths is a substantial evaluation effort, and the detailed description of the methods, including the KNS algorithm with background-target separation, is a strength. The authors are transparent about the simulation-based nature of the study and about the need for real-world validation. However, the evaluation protocol as reported prevents accepting the quantitative headline claims: hyperparameters are selected per plume by optimizing the very metric being reported, and identification performance is measured only by softmax confidence rather than by actual classification accuracy.

major comments (2)
  1. [IV-C] The central quantitative claims are computed under per-plume oracle hyperparameter selection. Section IV-C states that 'All of the previous figures illustrate results using the optimal hyperparameter for each plume' and that for identification confidence the search 'finds the hyperparameter that produces the largest confidence.' Because the true gas and true background are known in simulation, the hyperparameters are selected by exploiting the test plume's ground truth. Consequently, Tables I and II and Figures 4-8 report fitted maxima rather than predictive performance. The headline 53.2 percentage-point confidence gain and the 18,855x MSE improvement are upper bounds, and no fixed-hyperparameter or nested cross-validation result is reported. Please report results under a fixed configuration (e.g., the modes in Table III) or under a validation-based selection rule that does not use the test plume's label, and state clearly which numbers are predictive.
  2. [III-G / IV-B] The identification claim rests solely on softmax confidence for the true gas; classification accuracy (top-1 or top-k) and calibration are never reported. Section III-G defines the softmax output as 'identification confidence,' and Section IV-B uses only this quantity as the identification metric. A large increase in true-class confidence does not establish that the network identifies the gas more often, especially for a network trained and tested on the same simulation model where it may be overconfident. Please report top-1/top-k accuracy and, ideally, a reliability diagram or expected calibration error, so that the 'improved gas plume identification' claim in the title and abstract is directly supported.
minor comments (5)
  1. [III-F] The notation for clean versus contaminated pixels in the KNS description appears identical in the text ('Let Li represent clean pixels from Sn, and Li represent contaminated pixels from Sr'); please distinguish them, for example with superscripts or different subscripts.
  2. [IV-C] The text says 'a grid search of hyperparameters was tested' but then refers to the Optuna package, which is a hyperparameter optimization framework rather than an exhaustive grid search; please align the terminology.
  3. [IV-C] The phrase 'the number of minimum number of pixels found for values 2n with n in [2,11]' should be 'the minimum number of pixels'; this appears to be a typo.
  4. [Table I] The units of MSE are not specified; please state that MSE is in squared radiance units (or define the normalization used), so that the absolute values in Table I and Figure 4 are interpretable.
  5. [Abstract] The phrase '18,000 times less MSE' would be clearer as '18,000 times lower MSE' or 'a factor of 18,000 lower MSE' to avoid the awkward comparative.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline 53.2% confidence gain is computed under per-plume oracle hyperparameter selection that maximizes that same confidence value.

  1. fitted input called prediction [Section IV-B (Identification Confidence) and Section IV-C (Hyperparameter Selection), Tables II-III]
    "For each method and each plume, a grid search of hyperparameters was tested to determine which results in the best performance [54]. For background estimation, the search finds the hyperparameter that produces the smallest MSE, while for identification confidence, the search finds the hyperparameter that produces the largest confidence. ... All of the previous figures illustrate results using the optimal hyperparameter for each plume."

    The central claim that KNS raises median identification confidence from 38.2% to 91.4% is obtained by selecting, for each plume individually, the KNS hyperparameters (k, linkage, and whether to use BTS) that maximize that plume's identification confidence. Because the true gas is known in the simulation, this is oracle tuning: the reported 91.4% confidence is the maximum over the searched hyperparameter grid, not the output of a fixed method. Table III confirms that the optimal settings differ between tasks and plumes (e.g., KNS mode is k=4, single linkage, no BTS for identification but k=2048, average linkage, BTS for background estimation), and no fixed-hyperparameter configuration is reported.

full rationale

The only clear circularity pattern in this paper is the per-plume hyperparameter selection that directly maximizes the reported identification-confidence metric. The paper is transparent about this ('All of the previous figures illustrate results using the optimal hyperparameter for each plume'), but transparency does not remove the reduction: the abstract's 53.2% improvement and Table II's 91.4% median are maxima over per-plume grid searches, not fixed-algorithm predictions. This is a textbook fitted-input-called-prediction case, so the score is a 6 rather than a 0. I do not find the self-citations to be load-bearing circularity: the NNI is extended from prior work by overlapping authors, but it serves as a fixed classifier and the paper's contribution is the background estimators, not a derivation of the network. The use of the same forward model for training signatures and simulated test plumes is a closed-loop evaluation concern, but it is a validity/correctness risk rather than a derivation that reduces to its own inputs. No uniqueness theorem or ansatz is smuggled in through self-citation, and no known result is merely renamed. The core methodological contributions (KNN and KNS background estimators, the comparison across 640 simulated plumes) have independent empirical content; the circularity is concentrated in how the headline performance numbers are selected and reported.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

The central claim depends on the per-plume tuned hyperparameters, which are free parameters fitted to the evaluation data. The axioms are the physical and statistical simplifications that make the simulation and evaluation tractable. No new physical entities are introduced.

free parameters (8)
  • PCA number of components = 1-127 per plume (mode 127 for MSE, mode 1 for confidence)
    Selected per plume by Optuna grid search on the evaluation metric; Tables I-II use these oracle values.
  • KNN number of neighbors k = 1-127 per plume (mode 6/9 for MSE, mode 1 for confidence)
    Selected per plume by grid search on the target metric.
  • KMeans number of clusters k = 2-128 per plume (mode 128/124 for MSE, mode 2 for confidence)
    Selected per plume by grid search on the target metric.
  • Annulus number of dilations k = 1-127 per plume (mode 1/10 for MSE, mode 1 for confidence)
    Selected per plume by grid search on the target metric.
  • KNS minimum pixels k = 2^n, n in [2,11], per plume
    Selected per plume together with linkage function and BTS flag.
  • KNS linkage function = single, complete, or average per plume
    Chosen per plume; average linkage preferred for MSE, single for confidence.
  • KNS use of background-target separation = True or False per plume
    Binary hyperparameter tuned per plume; BTS preferred for MSE but not for confidence.
  • Simulation nc calibration = chosen per gas and signal strength category to hit 10%-80% ACE true positive rate
    The plume concentration-pathlength is calibrated so simulated signals match target detection rates, affecting the realism of the benchmark.
assumptions (6)
  • domain assumption Additive plume model L = L_off + psi*s (Eq. 1) is an adequate approximation of the full radiative transfer equation (Eq. 4).
    Used throughout to define background estimation, to generate training signatures (Eq. 12), and to simulate test plumes. The paper itself notes psi should be a vector and that background terms appear in both L_off and psi.
  • domain assumption Pixels outside the dilated ROI (set N) are uncontaminated by the plume (psi=0).
    All methods estimate the background from N; a four-dilation guardrail is assumed sufficient to exclude weak signal leakage.
  • domain assumption Softmax confidence of the neural network for the true gas class is a valid measure of identification performance.
    The paper's central result is a confidence increase; no accuracy or calibration validation is provided.
  • domain assumption The 640 simulated plumes, generated from 10 LA Basin images with estimated atmospheric and temperature-emissivity parameters, are representative of real operational plume scenes.
    The authors state in the conclusion that real-world validation is needed.
  • domain assumption The neural network trained on dynamically simulated signatures (Eq. 12) generalizes to the simulated test plumes in a way that reflects real-world identification performance.
    Training and test data come from the same simulation pipeline, so in-distribution performance may be optimistic.
  • standard math Standard PCA, K-means, watershed segmentation, and L-BFGS-B optimization behave as expected on hyperspectral data.
    Used as building blocks; no new theory is required.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improved Background Estimation for Gas Plume Identification in Hyperspectral Images." pith.science (2026). https://pith.science/paper/KY2HXRPT

@misc{pith2026241115378,
  author       = {Pith},
  title        = {Pith review of: Improved Background Estimation for Gas Plume Identification in Hyperspectral Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KY2HXRPT}},
  note         = {Machine review of arXiv:2411.15378}
}
read the original abstract

Longwave infrared (LWIR) hyperspectral imaging can be used for many tasks in remote sensing, including detecting and identifying effluent gases by LWIR sensors on airborne platforms. Once a potential plume has been detected, it needs to be identified to determine exactly what gas or gases are present in the plume. During identification, the background underneath the plume needs to be estimated and removed to reveal the spectral characteristics of the gas of interest. Current standard practice is to use ``global" background estimation, where the average of all non-plume pixels is used to estimate the background for each pixel in the plume. However, if this global background estimate does not model the true background under the plume well, then the resulting signal can be difficult to identify correctly. The importance of proper background estimation increases when dealing with weak signals, large libraries of gases of interest, and with uncommon or heterogeneous backgrounds. In this paper, we propose two methods of background estimation, in addition to three existing methods, and compare each against global background estimation to determine which perform best at estimating the true background radiance under a plume, and for increasing identification confidence using a neural network classification model. We compare the different methods using 640 simulated plumes. We find that PCA is best at estimating the true background under a plume, with a median of 18,000 times less MSE compared to global background estimation. Our proposed K-Nearest Segments algorithm improves median neural network identification confidence by 53.2%.

Figures

Figures reproduced from arXiv: 2411.15378 by the authors.

Figure 1
Figure 1. Example of gas plume identification process: 1. False color LWIR image from the Los Angeles Basin with a simulated N [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Diagram of the NNI architecture. The input is a 1-D vector, which [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 4
Figure 4. Background estimation MSE aggregated across all 640 simulated [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: Plot of the distribution of MSE scores for each method when [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Left: Distribution of NNI confidence across all 640 plumes and each [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Distribution of NNI confidence by gas when using Global background [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Identification confidence for hard to identify gases C [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Optimal choice of the number of components for PCA for both [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Hyperparameter sensitivity for each method. The right plot shows [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 56 canonical work pages

  1. [1]

    Hyperspectral imaging in environmental monitoring: A review of recent developments and technological advances in compact field deployable systems,

    M. B. Stuart, A. J. S. McGonigle, and J. R. Willmott, “Hyperspectral imaging in environmental monitoring: A review of recent developments and technological advances in compact field deployable systems,” Sensors (Basel, Switzerland) , vol. 19, 2019. [Online]. Available: https://api.semanticscholar.org/CorpusID:198192705

    work page 2019

  2. [2]

    New perspectives of hyperspectral imaging for clinical research,

    J. D. Pallua, A. Brunner, B. Zelger, C. W. Huck, M. Schirmer, J. Laimer, D. Putzer, M. Thaler, and B. Zelger, “New perspectives of hyperspectral imaging for clinical research,” NIR News, vol. 32, pp. 5 – 13, 2021. [Online]. Available: https://api.semanticscholar.org/CorpusID: 235760916

    work page 2021

  3. [3]

    An introduction to hyperspectral imaging and its application for security, surveillance and target acquisition,

    P. W. T. Yuen and M. A. Richardson, “An introduction to hyperspectral imaging and its application for security, surveillance and target acquisition,” The Imaging Science Journal, vol. 58, pp. 241 – 253, 2010. [Online]. Available: https://api.semanticscholar.org/CorpusID:34821906

    work page 2010

  4. [4]

    Modern trends in hyperspectral image analysis: A review,

    M. J. Khan, H. S. Khan, A. Yousaf, K. Khurshid, and A. Abbas, “Modern trends in hyperspectral image analysis: A review,” IEEE Access, vol. 6, pp. 14 118–14 129, 2018. [Online]. Available: https: //api.semanticscholar.org/CorpusID:4624178

    work page 2018

  5. [5]

    High spatial resolution imaging of methane and other trace gases with the airborne hyperspectral thermal emission spectrometer (hytes),

    G. C. Hulley, R. M. Duren, F. M. Hopkins, S. J. Hook, N. Vance, P. C. Guillevic, W. R. Johnson, B. T. Eng, J. M. Mihaly, V . M. Jovanovic, S. L. Chazanoff, Z. K. Staniszewski, L. Kuai, J. R. Worden, C. Frankenberg, G. Rivera, A. D. Aubrey, C. E. Miller, N. K. Malakar, J. M. S. Tom ´as, and K. T. Holmes, “High spatial resolution imaging of methane and othe...

    work page 2016

  6. [6]

    Overview of physical models and statistical approaches for weak gaseous plume detection using passive infrared hyperspectral imagery,

    T. L. Burr and N. W. Hengartner, “Overview of physical models and statistical approaches for weak gaseous plume detection using passive infrared hyperspectral imagery,” Sensors, vol. 6, pp. 1721–1750, 2006. [Online]. Available: https://api.semanticscholar.org/CorpusID:6554182

    work page 2006

  7. [7]

    Detection algorithms in hyperspectral imaging systems: An overview of practical algorithms,

    D. G. Manolakis, E. Truslow, M. L. Pieper, T. W. Cooley, and M. Brueggeman, “Detection algorithms in hyperspectral imaging systems: An overview of practical algorithms,” IEEE Signal Processing Magazine, vol. 31, pp. 24–33, 2014. [Online]. Available: https: //api.semanticscholar.org/CorpusID:18877994

    work page 2014

  8. [8]

    Detection algorithms for hyperspectral imaging applications,

    D. G. Manolakis and G. A. Shaw, “Detection algorithms for hyperspectral imaging applications,” IEEE Signal Process. Mag., vol. 19, pp. 29–43, 2002. [Online]. Available: https://api.semanticscholar.org/ CorpusID:62194350

    work page 2002

Show all 58 references
  1. [9]

    Performance metrics for the evaluation of hyperspectral chemical identification systems,

    E. Truslow, S. E. Golowich, D. G. Manolakis, and V . K. Ingle, “Performance metrics for the evaluation of hyperspectral chemical identification systems,” Optical Engineering , vol. 55, 2016. [Online]. Available: https://api.semanticscholar.org/CorpusID:53578703

    2016

  2. [10]

    Chemical identification using bayesian model selection,

    T. L. Burr, H. Fry, B. McVey, and E. Sander, “Chemical identification using bayesian model selection,” 2002. [Online]. Available: https://api.semanticscholar.org/CorpusID:17843691

    2002

  3. [11]

    Bayesian model averaging: A tutorial,

    J. A. Hoeting, D. Madigan, A. E. Raftery, and C. V olinsky, “Bayesian model averaging: A tutorial,” 1999. [Online]. Available: https://api.semanticscholar.org/CorpusID:60698623

    1999

  4. [12]

    Gas plume species identification in airborne lwir imagery using constrained stepwise regression analyses,

    D. R. Pogorzala, D. W. Messinger, C. Salvaggio, and J. R. Schott, “Gas plume species identification in airborne lwir imagery using constrained stepwise regression analyses,” 2005. [Online]. Available: https://api.semanticscholar.org/CorpusID:11211080

    2005

  5. [13]

    Deep learning for hyperspectral image classification: An overview,

    S. Li, W. Song, L. Fang, Y . Chen, P. Ghamisi, and J. A. Benediktsson, “Deep learning for hyperspectral image classification: An overview,” IEEE Transactions on Geoscience and Remote Sensing, vol. 57, pp. 6690–6709, 2019. [Online]. Available: https: //api.semanticscholar.org/C...

    2019

  6. [14]

    Hyperspectral target identification using physics- guided neural networks with explainability and feature attribution,

    N. Klein, A. Carr, Z. Hampel-Arias, A. Zastrow, A. Ziemann, and E. Flynn, “Hyperspectral target identification using physics- guided neural networks with explainability and feature attribution,” IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium , pp...

    2023

  7. [15]

    Background radiance estimation for gas plume quantification for airborne hyperspectral thermal imaging,

    R. Idoughi, T. H. G. Vidal, P.-Y . Foucher, M. A. Gagnon, and X. Briottet, “Background radiance estimation for gas plume quantification for airborne hyperspectral thermal imaging,” Spectroscopy, vol. 2016, pp. 1– 17, 2016. [Online]. Available: https://api.semanticscholar.org/C...

    2016

  8. [16]

    Remote trace gas quantification using thermal ir spectroscopy and digital filtering based on principal components of background scene clutter,

    A. F. Hayden and R. J. Noll, “Remote trace gas quantification using thermal ir spectroscopy and digital filtering based on principal components of background scene clutter,” in Defense, Security, and Sensing , 1997. [Online]. Available: https://api.semanticscholar.org/ CorpusI...

    1997

  9. [17]

    Longwave infrared hyperspectral imaging: Principles, progress, and challenges,

    D. G. Manolakis, M. L. Pieper, E. Truslow, R. B. Lockwood, A. Weisner, J. Jacobson, and T. W. Cooley, “Longwave infrared hyperspectral imaging: Principles, progress, and challenges,” IEEE Geoscience and Remote Sensing Magazine , vol. 7, pp. 72–100, 2019. [Online]. Available: h...

    2019

  10. [18]

    Long- wave infrared hyperspectral remote sensing of chemical clouds: A focus on signal processing approaches,

    D. G. Manolakis, S. E. Golowich, and R. S. DiPietro, “Long- wave infrared hyperspectral remote sensing of chemical clouds: A focus on signal processing approaches,” IEEE Signal Processing Magazine, vol. 31, pp. 120–141, 2014. [Online]. Available: https: //api.semanticscholar.o...

    2014

  11. [19]

    Tracking and quantification of gaseous chemical plumes from anthropogenic emission sources within the los angeles basin,

    K. N. Buckland, S. J. Young, E. R. Keim, B. R. Johnson, P. D. Johnson, and D. M. Tratt, “Tracking and quantification of gaseous chemical plumes from anthropogenic emission sources within the los angeles basin,” Remote Sensing of Environment , vol. 201, pp. 275–296,

  12. [20]

    Hyperspectral target detection : An overview of current and future challenges,

    N. M. Nasrabadi, “Hyperspectral target detection : An overview of current and future challenges,” IEEE Signal Processing Magazine, vol. 31, pp. 34–44, 2014. [Online]. Available: https://api.semanticscholar.org/CorpusID:12240632

    2014

  13. [21]

    An overview of background modeling for detection of targets and anomalies in hyperspectral remotely sensed imagery,

    S. Matteoli, M. Diani, and J. Theiler, “An overview of background modeling for detection of targets and anomalies in hyperspectral remotely sensed imagery,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 7, pp. 2317–2336, 2014. [Online]...

    2014

  14. [22]

    Is there a best hyperspectral detection algorithm?

    D. G. Manolakis, R. B. Lockwood, T. W. Cooley, and J. Jacobson, “Is there a best hyperspectral detection algorithm?” in Defense + Commercial Sensing , 2009. [Online]. Available: https://api.semanticscholar.org/CorpusID:26987112

    2009

  15. [23]

    A cluster-based approach for detecting man-made objects and changes in imagery,

    M. J. Carlotto, “A cluster-based approach for detecting man-made objects and changes in imagery,” IEEE Transactions on Geoscience and Remote Sensing , vol. 43, pp. 374–387, 2005. [Online]. Available: https://api.semanticscholar.org/CorpusID:11003912

    2005

  16. [24]

    Using elliptically contoured distributions to model hyperspectral imaging data and generate statistically similar synthetic data,

    D. Marden and D. G. Manolakis, “Using elliptically contoured distributions to model hyperspectral imaging data and generate statistically similar synthetic data,” in SPIE Defense + Commercial Sensing, 2004. [Online]. Available: https://api.semanticscholar.org/ CorpusID:122244177

    2004

  17. [25]

    Models and methods for automated background density estimation in hyperspectral anomaly detection,

    S. Matteoli, T. Veracini, M. Diani, and G. Corsini, “Models and methods for automated background density estimation in hyperspectral anomaly detection,” IEEE Transactions on Geoscience and Remote Sensing, vol. 51, pp. 2837–2852, 2013. [Online]. Available: https://api.semantics...

    2013

  18. [26]

    Ensemble segmentation for improved background estimation and gas plume identification in hyperspectral images,

    S. Jarman, T. Carr, Z. Hampel-Arias, E. Flynn, and K. R. Moon, “Ensemble segmentation for improved background estimation and gas plume identification in hyperspectral images,” in Optical Engineering + Applications , 2023. [Online]. Available: https://api.semanticscholar. org/C...

    2023

  19. [27]

    A k-means clustering algorithm,

    J. A. Hartigan and M. A. Wong, “A k-means clustering algorithm,” 1979. [Online]. Available: https://api.semanticscholar.org/CorpusID:53880671

    1979

  20. [28]

    k-means++: the advantages of careful seeding,

    D. Arthur and S. Vassilvitskii, “k-means++: the advantages of careful seeding,” in ACM-SIAM Symposium on Discrete Algorithms , 2007. [Online]. Available: https://api.semanticscholar.org/CorpusID:1782131

    2007

  21. [29]

    Clustering to improve matched filter detection of weak gas plumes in hyperspectral thermal imagery,

    C. Funk, J. Theiler, D. A. Roberts, and C. Borel-Donohue, “Clustering to improve matched filter detection of weak gas plumes in hyperspectral thermal imagery,” IEEE Trans. Geosci. Remote. Sens. , vol. 39, pp. 1410–1420, 2001. [Online]. Available: https://api.semanticscholar.or...

    2001

  22. [30]

    Principal component analysis,

    I. T. Jolliffe, “Principal component analysis,” 2002. [Online]. Available: https://api.semanticscholar.org/CorpusID:265962117

    2002

  23. [31]

    Hyperspectral subpixel target detection using the linear mixing model,

    D. G. Manolakis, C. Siracusa, and G. A. Shaw, “Hyperspectral subpixel target detection using the linear mixing model,” IEEE Trans. Geosci. Remote. Sens. , vol. 39, pp. 1392–1409, 2001. [Online]. Available: https://api.semanticscholar.org/CorpusID:43627459

    2001

  24. [32]

    Determination of trace-gas amounts in plumes by the use of orthogonal digital filtering of thermal-emission spectra

    A. F. Hayden, E. R. Niple, and B. E. Boyce, “Determination of trace-gas amounts in plumes by the use of orthogonal digital filtering of thermal-emission spectra.” Applied optics, vol. 35 16, pp. 2802–9, 1996. [Online]. Available: https://api.semanticscholar.org/CorpusID:985274

    1996

  25. [33]

    Introduction to machine learning: k-nearest neighbors

    W. Bao, “Introduction to machine learning: k-nearest neighbors.” Annals of translational medicine , vol. 4 11, p. 218, 2016. [Online]. Available: https://api.semanticscholar.org/CorpusID:42966081

    2016

  26. [34]

    Local manifold learning- based $k$ -nearest-neighbor for hyperspectral image classification,

    L. Ma, M. M. Crawford, and J. Tian, “Local manifold learning- based $k$ -nearest-neighbor for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 48, pp. 4099–4109, 2010. [Online]. Available: https://api.semanticscholar.org/ CorpusID:29808774

    2010

  27. [35]

    Spectral–spatial hyperspectral image classification based on knn,

    K. Huang, S. Li, X. Kang, and L. Fang, “Spectral–spatial hyperspectral image classification based on knn,” Sensing and Imaging , vol. 17,

  28. [36]

    Class-dependent sparse representation classifier for robust hyperspectral image classification,

    M. Cui and S. Prasad, “Class-dependent sparse representation classifier for robust hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 53, pp. 2683–2695, 2015. [Online]. Available: https://api.semanticscholar.org/CorpusID:5214904

    2015

  29. [37]

    K-nearest neighbor combined with guided filter for hyperspectral image classification,

    Y . Guo, S. Han, Y . Li, C. Zhang, and Y . Bai, “K-nearest neighbor combined with guided filter for hyperspectral image classification,” in International Conference on Identification, Information, and Knowledge in the Internet of Things , 2017. [Online]. Available: https://api...

    2017

  30. [38]

    Background density nonparametric estimation with data-adaptive bandwidths for the detection of anomalies in multi-hyperspectral imagery,

    S. Matteoli, T. Veracini, M. Diani, and G. Corsini, “Background density nonparametric estimation with data-adaptive bandwidths for the detection of anomalies in multi-hyperspectral imagery,” IEEE Geoscience and Remote Sensing Letters , vol. 11, pp. 163–167, 2014. [Online]. Ava...

    2014

  31. [39]

    A locally adaptive background density estimator: An evolution for rx-based anomaly detectors,

    ——, “A locally adaptive background density estimator: An evolution for rx-based anomaly detectors,” IEEE Geoscience and Remote Sensing Letters , vol. 11, pp. 323–327, 2014. [Online]. Available: https://api.semanticscholar.org/CorpusID:17278403

    2014

  32. [40]

    Adaptive detection algorithm for full pixel targets in hyperspectral images,

    N. Acito, G. Corsini, and M. Diani, “Adaptive detection algorithm for full pixel targets in hyperspectral images,” 2005. [Online]. Available: https://api.semanticscholar.org/CorpusID:140141650

    2005

  33. [41]

    A tutorial overview of anomaly detection in hyperspectral images,

    S. Matteoli, M. Diani, and G. Corsini, “A tutorial overview of anomaly detection in hyperspectral images,” IEEE Aerospace and Electronic Systems Magazine , vol. 25, pp. 5–28, 2010. [Online]. Available: https://api.semanticscholar.org/CorpusID:36740158

    2010

  34. [42]

    Hyperspectral image segmentation: a comprehensive survey,

    R. Grewal, S. S. Kasana, and G. Kasana, “Hyperspectral image segmentation: a comprehensive survey,” Multimedia Tools and Applications, vol. 82, pp. 20 819 – 20 872, 2022. [Online]. Available: https://api.semanticscholar.org/CorpusID:253231018

    2022

  35. [43]

    Segmentation and classification of hyperspectral images using watershed transformation,

    Y . Tarabalka, J. Chanussot, and J. A. Benediktsson, “Segmentation and classification of hyperspectral images using watershed transformation,” Pattern Recognit., vol. 43, pp. 2367–2379, 2010. [Online]. Available: https://api.semanticscholar.org/CorpusID:26615105

    2010

  36. [44]

    Cluster merging and splitting in hierarchical clustering algorithms,

    C. Ding and X. He, “Cluster merging and splitting in hierarchical clustering algorithms,” 2002 IEEE International Conference on Data Mining, 2002. Proceedings. , pp. 139–146, 2002. [Online]. Available: https://api.semanticscholar.org/CorpusID:7563771

    2002

  37. [45]

    A comprehensive evaluation of spectral distance functions and metrics for hyperspectral image processing,

    H. Deborah, N. Richard, and J. Y . Hardeberg, “A comprehensive evaluation of spectral distance functions and metrics for hyperspectral image processing,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 8, pp. 3224–3234, 2015. [Online]. A...

    2015

  38. [46]

    Local background estimation for improved gas plume identification in hyper- spectral images,

    S. Jarman, Z. Hampel-Arias, A. Carr, and K. R. Moon, “Local background estimation for improved gas plume identification in hyper- spectral images,” IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium , pp. 8005–8009, 2024. [Online]. Available: https:/...

    2024

  39. [47]

    A limited memory algorithm for bound constrained optimization,

    R. H. Byrd, P. Lu, J. Nocedal, and C. Zhu, “A limited memory algorithm for bound constrained optimization,” SIAM J. Sci. Comput. , vol. 16, pp. 1190–1208, 1995. [Online]. Available: https://api.semanticscholar. org/CorpusID:6398414

    1995

  40. [48]

    Quantifying uncertainty in machine learning for hyperspectral target detection and identification,

    N. Klein, A. Carr, Z. Hampel-Arias, A. Ziemann, E. Flynn, and K. Mitchell, “Quantifying uncertainty in machine learning for hyperspectral target detection and identification,” in Defense + Commercial Sensing , 2022. [Online]. Available: https://api.semanticscholar.org/CorpusID...

    2022

  41. [49]

    2d spectral representations and autoencoders for hyperspectral imagery classification and explanability,

    Z. Hampel-Arias, A. Carr, N. Klein, and E. Flynn, “2d spectral representations and autoencoders for hyperspectral imagery classification and explanability,” 2024 IEEE Southwest Symposium on Image Analysis and Interpretation (SSIAI) , pp. 45–48, 2024. [Online]. Available: https...

    2024

  42. [50]

    Lets keep it simple, using simple architectures to outperform deeper and more IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. XX, 202X 13 complex architectures,

    S. H. HasanPour, M. Rouhani, M. Fayyaz, and M. Sabokrou, “Lets keep it simple, using simple architectures to outperform deeper and more IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. XX, 202X 13 complex architectures,” ArXiv, vol. abs/1608.06037, 2016. [Online]. Avai...

    2016 arXiv

  43. [51]

    Squeeze-and- excitation networks,

    J. Hu, L. Shen, S. Albanie, G. Sun, and E. Wu, “Squeeze-and- excitation networks,” 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 7132–7141, 2017. [Online]. Available: https://api.semanticscholar.org/CorpusID:140309863

    2018

  44. [52]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” CoRR, vol. abs/1412.6980, 2014. [Online]. Available: https://api.semanticscholar.org/CorpusID:6628106

    2014 arXiv

  45. [53]

    A review of dispersion modelling and its application to the dispersion of particles : An overview of different dispersion models available,

    N. S. Holmes and L. Morawska, “A review of dispersion modelling and its application to the dispersion of particles : An overview of different dispersion models available,” Atmospheric Environment , vol. 40, pp. 5902–5928, 2006. [Online]. Available: https://api.semanticscholar....

    2006

  46. [54]

    Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges,

    B. Bischl, M. Binder, M. Lang, T. Pielok, J. Richter, S. Coors, J. Thomas, T. Ullmann, M. Becker, A. Boulesteix, D. Deng, and M. T. Lindauer, “Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges,” Wiley Interdisciplinary Reviews: Data Mini...

    2021

  47. [55]

    Optuna: A next-generation hyperparameter optimization framework,

    T. Akiba, S. Sano, T. Yanase, T. Ohta, and M. Koyama, “Optuna: A next-generation hyperparameter optimization framework,” Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining , 2019. [Online]. Available: https://api. semanticscholar.o...

    2019

  48. [2015]

    Available: https://api.semanticscholar.org/CorpusID: 123815316

    [Online]. Available: https://api.semanticscholar.org/CorpusID: 123815316

  49. [2017]

    Available: https://api.semanticscholar.org/CorpusID: 134128864

    [Online]. Available: https://api.semanticscholar.org/CorpusID: 134128864

  50. [2018]

    His re- search interests are in the development of theory and applications in machine learning, information theory, deep learning, and manifold learning

    in the Genetics Department and the Applied Mathematics Program at Yale University. His re- search interests are in the development of theory and applications in machine learning, information theory, deep learning, and manifold learning. More specifically, he is interested in d...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.