Pith.

Computer Vision and Pattern Recognition · arXiv:2611.17199 v1 · submitted 12 Nov 2026 · 49 references

MALNISI-Infused Graph Construction and Weighted Contrastive Loss based GAT for Unsupervised Representation Learning in HSI

Soumi Sarkar, Subhadip Boral, Sarbani Palit, Ashish Ghosh

The paper says

“Hyperspectral images (HSI) possess discriminating spectral characteristics with respect to land utilization due to high-dimensional information. Recently, unsupervised representation learning using graph neural networks has unveiled these utilization-specific complex spectral relations. This article has three main contributions.”

more of the abstract

First, it proposes an index, the Mean Absolute Local Neighborhood Inverse Spectral Index (MALNISI), which measures the similarity between two pixels based on their neighborhood information. The index is used to construct a weighted adjacency matrix, giving the HSI a graph structure. Second, it proposes an upgraded Graph Attention Network that learns compact latent representations of hyperspectral data from the constructed graph. In this network, an enhanced multi-head attention mechanism incorporates the MALNISI weights as an auxiliary feature to emphasize learned attention between neighboring pixels. To alleviate over-smoothing and preserve fine-grained spectral details, the architecture integrates residual connections, skip projections, and feature normalization, which together enhance representation capability, stabilize learning, and accelerate convergence. Third, it proposes a weighted contrastive loss function that incorporates neighborhood-aware MALNISI between connected pixels, encouraging strongly connected pixels in the high-dimensional space to remain close in the latent space, and weakly connected pixels to stay apart. The resulting discriminative low-dimensional embeddings are then clustered using spectral clustering. Experimental analysis on six benchmark HSI datasets demonstrates the efficiency and necessity of the three aforementioned contributions, and the comparative study highlights superior performance over fourteen state-of-the-art approaches.

from the abstract

Pith says

Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded · not checked by a person

It can be wrong. Challenge this review or reply to it.

Holds up, with conditions. MALNISI is a plausible new similarity index for HSI graphs, but the paper has a load-bearing formula inconsistency that needs resolution before the results can be trusted. This line is wrong

5 major objections · 12 minor

How often objections from reviews like this one are right: not yet measured

Are you an author? What you can do.

Not checked: figures, code, what the 49 references say.

The paper says, at §III-A, Eq. (3)

The review did not quote this passage.

Pith says: Objection 1 of 5 · major

The text in §III-A states that 'the normalization factor 1/|N_{u,v}| makes MALNISI unbiased towards the neighborhood function,' but Equation (3) prints MALNISI(u,v) = 1 / (1 + d^{-1/2} |N_{u,v}| SI(u,v)), which multiplies by |N_{u,v}| instead of dividing. This is a load-bearing inconsistency because the claimed invariance to neighborhood size depends on the correct normalization.

more of objection 1

The paper does not clarify which version was implemented. If Equation (3) is literal, every MALNISI edge weight is miscalibrated with respect to neighborhood size, and the reported comparisons may not reflect the published index. The authors must resolve this discrepancy by stating explicitly which formula was coded and, if the text version is correct, correcting Equation (3).

The paper says, at §V-B, Table VII

The review did not quote this passage.

Pith says: Objection 2 of 5 · major

Table VII reports standard deviations for the proposed method but not for any of the fourteen baseline methods. Without baseline standard deviations, the t-test p-values reported in the same table are uninterpretable because the test requires variance estimates for both groups.

more of objection 2

The authors should either provide standard deviations for the baselines (e.g., from multiple runs or from the original papers) or clarify the source of the baseline variances used in the t-test. As it stands, the statistical significance claims are not supported.

The paper says, at §V-A, Table VI and sentence [b182]

The review did not quote this passage.

Pith says: Objection 3 of 5 · major

The text states 'Proposed work achieves identical NMI and AMI values across all datasets,' but Table VI shows NMI and AMI values that are close but not identical (e.g., IndianP: NMI=0.7818, AMI=0.7811; PaviaU: NMI=0.9206, AMI=0.9205; Salinas: NMI=0.9279, AMI=0.9279; HongHu: NMI=0.7964, AMI=0.7963; Botswana: NMI=0.8214, AMI=0.8198; KSC: NMI=0.885, AMI=0.8844).

more of objection 3

The claim of 'identical' is false; the values differ at the third or fourth decimal place. The authors should correct this overstatement and explain why the values are nearly equal (which is plausible for well-separated clusters) rather than claiming exact identity.

This review says the paper's mathematics is wrong. Pith did not check that against the paper's source.

The paper says, at §V-B, Table VII, Salinas row

The review did not quote this passage.

Pith says: Objection 4 of 5 · major

The proposed method's OA on Salinas (86.69) is slightly lower than the best baseline EGFSC (87.28), and the t-statistic is negative (-12.68) with p-value 0.9999, indicating the proposed method is not significantly better. The text acknowledges this ('highly competitive OA performance (−0.04%)') but the framing in the abstract and conclusion emphasizes 'superior performance over fourteen state-of-the-art approaches' without qualification.

more of objection 4

The authors should explicitly state that on Salinas the method is competitive but not superior in OA, and discuss why this dataset behaves differently (e.g., high spectral separability reduces the need for the proposed components).

The paper says, at §V-B, Table VIII, Botswana

The review did not quote this passage.

Pith says: Objection 5 of 5 · major

The proposed method is second-best on Botswana behind DSCFS, with OA 80.70 vs 82.39. The text acknowledges this but argues that DSCFS's performance drops on dense datasets. However, the paper does not provide evidence for this claim (e.g., DSCFS results on IndianP or PaviaU).

more of objection 5

The authors should either include DSCFS results on dense datasets from the same experimental setup or cite the original paper's reported performance on those datasets. Without this, the claim that the proposed method is 'superior' overall is weakened.

12 minor objections

The paper says, at §III-A, Eq. (2)

Pith says: minor objection 1 of 12

The notation in Eq. (2) uses double vertical bars for the norm, which is fine, but the absolute value signs around the difference of norms are written as ‖‖‖‖, which may be a formatting artifact. Please ensure the equation is typeset correctly.

The paper says, at §IV-C, Eq. (5)

Pith says: minor objection 2 of 12

The notation N4(s_i) and N8(·) is introduced but not defined. It is clear from context that these refer to 4-neighborhood and 8-neighborhood in the spatial grid, but a brief definition would improve clarity.

show 10 more

The paper says, at §V-A, Table VI

Pith says: minor objection 3 of 12

The table includes a column 'Dunn' with a dash for HongHu, and the text explains that the Dunn index could not be computed due to computational cost. This is acceptable, but the authors should note that the dash indicates 'not computed' rather than a zero or missing value.

The paper says, at §V-B, Table VII

Pith says: minor objection 4 of 12

The table uses blue and teal for best and second best, but the caption says 'BEST AND SECOND BEST PERFORMANCE REPRESENTED IN BLUE AND TEAL.' However, the table appears to use blue for the proposed method (which is best in most cases) and teal for the best baseline. This is fine, but the caption should clarify that the proposed method is highlighted separately.

The paper says, at §V-C, Figure 5

Pith says: minor objection 5 of 12

The radar plots are informative, but the radial axes are not labeled with metric names or scales. The caption says 'the angular axes represent different evaluation metrics, the radial axes reflect their normalized range of values,' but the normalization is not explained. Please add a note on how normalization was performed (e.g., min-max scaling across methods for each metric).

The paper says, at §V-D, Figure 6

Pith says: minor objection 6 of 12

Similar comment as for Figure 5: the radar plots lack axis labels and normalization details.

The paper says, at §V-E, Figure 7

Pith says: minor objection 7 of 12

The noise sensitivity analysis uses NMI as the sole metric. While NMI is reasonable, the authors should justify why other metrics (e.g., OA, ARI) are not reported for noise robustness.

The paper says, at §V-F, Figure 8

Pith says: minor objection 8 of 12

The figure compares NMI of traditional and modern feature extraction methods. The bars for modern methods are hatched, but some datasets have missing bars (e.g., HongHu for some methods). The text notes this, but a table with the exact values would be more informative than a bar chart with gaps.

The paper says, at §VI

Pith says: minor objection 9 of 12

The conclusion mentions that the framework 'does not explicitly address sparse data' and 'class imbalance.' These are acknowledged limitations, which is good. However, the paper does not quantify the impact of class imbalance on the reported metrics (e.g., per-class accuracies). The authors could add a supplementary table with per-class F1 scores or recall.

The paper says, at General

Pith says: minor objection 10 of 12

The paper uses the term 'unsupervised representation learning' but the evaluation is clustering-based (spectral clustering on learned embeddings). This is standard for unsupervised HSI segmentation, but the distinction between representation learning and clustering should be clarified in the title or abstract to avoid confusion.

The paper says, at References

Pith says: minor objection 11 of 12

Reference [29] is a book by A. Ghosh published in 2026, which may not be widely available. Please verify that the citation is correct and that the book is indeed published or in press.

The paper says, at §II, Table I

Pith says: minor objection 12 of 12

The taxonomy table is useful, but some entries are vague (e.g., 'Fails to capture non-linear structures' for PCA/LDA). Consider adding more specific limitations or references.

What would settle it

Check the released code (the paper states 'The implementation of the proposed method is available in GitHub') for the actual MALNISI computation. If the code uses division by |N_{u,v}| (consistent with the text), the concern is resolved. If the code uses multiplication by |N_{u,v}| (consistent with Equation 3), then recompute Table VII with the corrected formula; if OA on IndianP drops by more than 2 percentage points, the headline claim weakens.

from the skeptic's reading

How the review went

4 readings by deepseek-v4-flash · 13 Nov 2026

  1. Reader holds up, with conditions confidence moderate
  2. Skeptic agrees 12 s · 21,140 tokens read, 716 written
  3. Editor wrote the verdict above 8 s · 22,108 tokens read, 613 written
  4. Referee major revision: 5 major objections, 12 minor 20 s · 21,654 tokens read, 2,096 written

Not run on this paper: the simulated authors' defence.

Pith did not check this review's objections against the paper.

The review in full

The review in brief, machine-written

The paper proposes a three-part unsupervised HSI segmentation pipeline: a new local-context similarity index (MALNISI), a GAT that incorporates MALNISI weights into attention, and a weighted contrastive loss that uses those same weights. The experimental scope is genuinely broad—six datasets, fourteen baselines, ablation studies on the index and the loss. The gains on Indian Pines and PaviaU are large enough to be interesting if they hold up.

What the paper does well: the idea of measuring pixel similarity via their joint neighborhood rather than pointwise spectra is sensible for HSI, where spatial context matters. The ablation studies (replacing MALNISI with other indices, replacing the weighted loss with standard alternatives) are the right experiments to isolate the contribution. The authors also acknowledge several limitations honestly—sensitivity to graph quality, computational cost, poor handling of sparse data and class imbalance—which is more than many papers do.

The soft spot is the one the stress-test flags: Equation (3) multiplies by |Nu,v| while the text says it divides. The paper's central claim that MALNISI is invariant to neighborhood size depends on the division version. If the literal equation was coded, every edge weight is miscalibrated with respect to neighborhood size, and the reported numbers may not reflect the published index. The paper itself does not clarify which was implemented. This is load-bearing because the entire pipeline—graph construction, attention reweighting, contrastive loss—depends on those weights. The authors need to state which formula was used and, ideally, release code to confirm.

Minor issues: the 'identical NMI and AMI' claim in the text contradicts Table VI (they are close but not identical). Baseline standard deviations are absent from Table VII, making it hard to assess significance beyond the t-tests they do report. The paper also claims O(N3) complexity from spectral clustering but then reports inference times that seem too fast for cubic scaling on the larger datasets—this deserves a brief explanation.

Who this is for: researchers working on unsupervised HSI segmentation or graph-based representation learning for remote sensing. The paper deserves a serious referee because the method is coherent and the experiments are extensive, but the referee must verify the formula discrepancy and request code release. I would not cite it until the formula issue is resolved.

Recommendation: send to peer review with a strong request to clarify Equation (3) and release code. If the authors confirm the text version and release code, the paper could be a solid contribution.

The strongest objection

The paper's central claim that MALNISI is invariant to neighborhood size depends on dividing SI by |N_{u,v}|, as stated in the text (Section III-A: 'the normalization factor 1/|N_{u,v}| makes MALNISI unbiased towards the neighborhood function'). However, Equation (3) prints MALNISI(u,v) = 1 / (1 + d^{-1/2} |N_{u,v}| SI(u,v)), which multiplies by |N_{u,v}| instead of dividing. If the literal equation was implemented, then for larger neighborhoods the denominator grows, making MALNISI smaller—exactly the opposite of the claimed invariance. The paper reports large gains on IndianP (OA 80.49 vs best baseline 70.08) and PaviaU (OA 90.61 vs 75.76), but if the implemented formula is the one in Equation (3), every edge weight is miscalibrated with respect to neighborhood size, and the reported comparisons may not reflect the published index. The authors acknowledge the discrepancy in the reader's verdict but do not resolve it; the paper itself does not clarify which version was coded. This is a load-bearing concern because the entire graph construction, attention reweighting, and weighted contrastive loss depend on the MALNISI values; if they are systematically wrong, the reported performance may be an artifact of a different (possibly fortuitous) weighting scheme rather than the claimed index.

The reader's weakest_assumption correctly identifies the formula mismatch between Equation (3) and the text. This is the single most load-bearing concern because the entire method—graph construction, attention reweighting, weighted contrastive loss—depends on MALNISI values. If the implemented formula is the one in Equation (3), every edge weight is miscalibrated with respect to neighborhood size, and the reported comparisons may not reflect the published index. The paper does not resolve this discrepancy; it does not state which version was coded. The concrete test is straightforward: check the released code. The verdict remains CONDITIONAL because the concern is resolvable by code inspection; if the code matches the text, the paper is acceptable. I agree with the reader that this is the weakest assumption. No other concern is as load-bearing: the paper acknowledges limitations (sparse data, class imbalance, computational cost) and the results are internally consistent. The formula mismatch is the one issue that could invalidate the central claim if resolved unfavorably.

The referee's report

If the results hold, the paper makes a solid contribution to unsupervised HSI segmentation by introducing a neighborhood-aware similarity index that is robust to noise and provides context-aware graph weights. The integration of this index into both graph construction and contrastive loss is a coherent design. The experimental evaluation is broad, covering six datasets and many baselines. However, the significance is tempered by the unresolved discrepancy in Equation (3) and the lack of standard deviations for baselines, which weaken the reliability of the claimed improvements.

Pith's reading of the paper (a machine's summary, not the authors')

Hyperspectral images record many light bands for every pixel, which lets algorithms tell crops, buildings, and vegetation apart. But labeling every pixel is expensive. This paper tries to group pixels automatically, with no labels, by building a graph: each pixel is a node, and nearby pixels are connected. The authors define a new way of measuring how similar two neighboring pixels are, called MALNISI. It compares how each pixel sees the shared pixels around it. If two pixels and their surroundings look spectrally consistent, they get a high edge weight; at boundaries between land types the weight drops. These weights form the graph that a graph attention network uses to compress the pixels into a few numbers per pixel while preserving neighborhood structure. A custom contrastive loss then pushes strongly connected pixels together in the compressed space and weakly connected or unconnected pixels apart.

Core claim. The proposed MALNISI-GAT framework, combining MALNISI graph weights, attention reweighting, and a weighted contrastive loss, outperforms fourteen state-of-the-art methods on six benchmark HSI datasets, with the largest gains on IndianP (OA 80.49 vs best baseline 70.08) and PaviaU (OA 90.61 vs 75.76), as reported in Table VII. The paper's own Table VIII shows the method is second-best on Botswana behind DSCFS.

The premise it rests on. MALNISI's claimed invariance to neighborhood size depends on dividing SI by |N_{u,v}|, as the text in Section III-A says, but Equation 3 prints multiplication by |N_{u,v}|. The experimental results require the text version to be the implemented formula; if Equation 3 is literal, every MALNISI edge weight is miscalibrated with respect to neighborhood size and the reported comparisons may not reflect the published index.

Check this review

Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.

How often reviews like this one are right

Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.

How often objections from reviews like this one are right: not yet measured

Every value above, in the review's public record (JSON)

Reviews by people

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review

How Pith checks papers · The full audit of this paper

The paper

Abstract

Hyperspectral images (HSI) possess discriminating spectral characteristics with respect to land utilization due to high-dimensional information. Recently, unsupervised representation learning using graph neural networks has unveiled these utilization-specific complex spectral relations. This article has three main contributions. First, it proposes an index, the Mean Absolute Local Neighborhood Inverse Spectral Index (MALNISI), which measures the similarity between two pixels based on their neighborhood information. The index is used to construct a weighted adjacency matrix, giving the HSI a graph structure. Second, it proposes an upgraded Graph Attention Network that learns compact latent representations of hyperspectral data from the constructed graph. In this network, an enhanced multi-head attention mechanism incorporates the MALNISI weights as an auxiliary feature to emphasize learned attention between neighboring pixels. To alleviate over-smoothing and preserve fine-grained spectral details, the architecture integrates residual connections, skip projections, and feature normalization, which together enhance representation capability, stabilize learning, and accelerate convergence. Third, it proposes a weighted contrastive loss function that incorporates neighborhood-aware MALNISI between connected pixels, encouraging strongly connected pixels in the high-dimensional space to remain close in the latent space, and weakly connected pixels to stay apart. The resulting discriminative low-dimensional embeddings are then clustered using spectral clustering. Experimental analysis on six benchmark HSI datasets demonstrates the efficiency and necessity of the three aforementioned contributions, and the comparative study highlights superior performance over fourteen state-of-the-art approaches.

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.

abstract compiled · 2026-11-14

The authors' main claim, as the classifier picked it out 1.00

“Experimental analysis on six benchmark HSI datasets demonstrates the efficiency and necessity of the three aforementioned contributions, and the comparative study highlights superior performance over fourteen state-of-the-art approaches.”

How the main claim is stated: asserted 1.00

What the abstract itself states, as probabilities: claims first 0.06 · states limitation 0.06 · human participants 0.02 · controlled comparison 0.56 · randomized assignment 0.01 · states quantitative result 0.46 · compares with prior results 0.97 · mentions code or data availability 0.02

The abstract, sentence by sentence

This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.

SentenceRoleThe authors' own workStates the main conclusionStrength
Hyperspectral images (HSI) possess discriminating spectral characteristics with respect to land utilization due to high-dimensional information. background 1.00 0.05 0.02 asserted 1.00
Recently, unsupervised representation learning using graph neural networks has unveiled these utilization-specific complex spectral relations. background 1.00 0.05 0.03 asserted 0.99
This article has three main contributions. contribution 0.99 0.47 0.02 asserted 0.91
First, it proposes an index, the Mean Absolute Local Neighborhood Inverse Spectral Index (MALNISI), which measures the similarity between two pixels based on their neighborhood information. contribution 1.00 0.94 0.03 asserted 1.00
The index is used to construct a weighted adjacency matrix, giving the HSI a graph structure. method 0.77 0.86 0.02 asserted 1.00
Second, it proposes an upgraded Graph Attention Network that learns compact latent representations of hyperspectral data from the constructed graph. contribution 0.99 0.92 0.04 asserted 1.00
In this network, an enhanced multi-head attention mechanism incorporates the MALNISI weights as an auxiliary feature to emphasize learned attention between neighboring pixels. method 0.79 0.90 0.03 asserted 1.00
To alleviate over-smoothing and preserve fine-grained spectral details, the architecture integrates residual connections, skip projections, and feature normalization, which together enhance representation capability, stabilize learning, and accelerate convergence. method 0.82 0.87 0.04 asserted 1.00
Third, it proposes a weighted contrastive loss function that incorporates neighborhood-aware MALNISI between connected pixels, encouraging strongly connected pixels in the high-dimensional space to remain close in the latent space, and weakly connected pixels to stay apart. contribution 0.99 0.93 0.08 asserted 0.99
The resulting discriminative low-dimensional embeddings are then clustered using spectral clustering. method 0.99 0.89 0.08 asserted 1.00
Experimental analysis on six benchmark HSI datasets demonstrates the efficiency and necessity of the three aforementioned contributions, and the comparative study highlights superior performance over fourteen state-of-the-art approaches. result 0.99 0.96 0.81 asserted 1.00

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.67

What was examined, and what was not

Sources read

  • arXiv listing abstract v1

Not examined

  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers

Levels run

  • P abstract

Not run

  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

How the labels were made

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-11-14: 1 request, 8,168 input tokens.

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON

Figures (7)
Figure 1 of the paper
Figure 1. Architectural overview of the proposed MALNISI-GAT framework,
Figure 2 of the paper
Figure 2. Illustration and analysis of the MALNISI-based graph construction,
Figure 3 of the paper
Figure 3. Detailed illustration of the proposed MALNISI-GAT architecture.
Figure 4 of the paper
Figure 4. Geometric visualization of the proposed weighted contrastive loss.
Figure 5 of the paper
Figure 5. Radar visualization of performance analysis of different similarity
Figure 6 of the paper
Figure 6. Radar visualization of performance analysis of different loss functions.
Figure 7 of the paper
Figure 7. Radar visualization of performance analysis under different noise

Figures are the authors', reproduced from the paper.

References (49)
  1. [1] C.-I. Chang,Hyperspectral imaging: techniques for spectral detection and classification. Springer Science & Business Media, 2003, vol. 1 on Pith
  2. [2] Y . Zhang, X. Wang, X. Jiang, L. Zhang, and B. Du, “Elastic graph fusion subspace clustering for large hyperspectral image,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 35, no. 7, pp. 6300– 6312, 2025 on Pith
  3. [3] G. Corso, H. Stark, S. Jegelka, T. Jaakkola, and R. Barzilay, “Graph neural networks,”Nature Reviews Methods Primers, vol. 4, no. 1, p. 17, 2024 on Pith
  4. [4] R. Wang, F. Nie, Z. Wang, F. He, and X. Li, “Scalable graph-based clustering with nonnegative relaxation for large hyperspectral image,” IEEE Transactions on Geoscience and Remote Sensing, vol. 57, no. 10, pp. 7352–7364, 2019 on Pith
  5. [5] S. Sarkar, S. Boral, S. Palit, and A. Ghosh, “Multi-branch local global fusion network for cyclone classification using domain aware pooling,” Information Sciences, vol. 741, p. 123258, 2026 on Pith
  6. [6] A. Polowczyk, A. Polowczyk, M. Wozniak, and M. Wieczorek, “Graph- Based Pixel Representation Using GCN for Semantic Face Segmenta- tion,” inIn Proceedings of the 18th International Conference on Agents and Artificial Intelligence. SCITEPRESS, 2026, pp. 901–909 on Pith
  7. [7] J. Boardman, “Spectral angle mapping: a rapid measure of spectral similarity,”AVIRIS. Delivered by Ingenta, 1993 on Pith
  8. [8] C.-I. Chang, “An information-theoretic approach to spectral variability, similarity, and discrimination for hyperspectral image analysis,”IEEE Transactions on information theory, vol. 46, no. 5, pp. 1927–1932, 2000 on Pith
  9. [9] H. A. Aky ¨urek and B. Koc ¸er, “Spectral similarity based multiscale spatial-spectral preprocessing framework for hyperspectral image clas- sification.”Traitement du Signal, vol. 41, no. 4, 2024 on Pith
  10. [10] C. Cariou, S. Le Moan, and K. Chehdi, “Improving k-nearest neighbor approaches for density-based pixel clustering in hyperspectral remote sensing images,”Remote Sensing, vol. 12, no. 22, p. 3745, 2020 on Pith
  11. [11] Y . Zhang, G. Jiang, Z. Cai, and Y . Zhou, “Bipartite graph-based pro- jected clustering with local region guidance for hyperspectral imagery,” IEEE Transactions on Multimedia, vol. 26, pp. 9551–9563, 2024 on Pith
  12. [12] G. Jiang, Y . Zhang, X. Wang, X. Jiang, and L. Zhang, “Structured anchor learning for large-scale hyperspectral image projected clustering,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 35, no. 3, pp. 2328–2340, 2024 on Pith
  13. [13] P. Veli ˇckovi´c, G. Cucurull, A. Casanova, A. Romero, P. Lio, and Y . Ben- gio, “Graph attention networks,”arXiv preprint arXiv:1710.10903, 2017 on PithPith's reviewarXiv
  14. [14] A. Ng, M. Jordan, and Y . Weiss, “On spectral clustering: Analysis and an algorithm,”Advances in Neural Information Processing Systems, vol. 14, 2001 on Pith
  15. [15] A. Datta, S. Ghosh, and A. Ghosh, “Unsupervised band extraction for hyperspectral images using clustering and kernel principal component analysis,”International Journal of Remote Sensing, vol. 38, no. 3, pp. 850–873, 2017 on Pith
  16. [16] K. R. Murthy and A. Ghosh, “Noisy-free length discriminant analysis with cosine hyperbolic framework for dimensionality reduction,”Expert Systems with Applications, vol. 81, pp. 88–107, 2017 on Pith
  17. [17] S. Boral, M. Sarkar, and A. Ghosh, “MEQA: Manifold embedding quality assessment via anisotropic scaling and kolmogorov-smirnov test,”Pattern Recognition, vol. 139, p. 109447, 2023 on Pith
  18. [18] C.-W. Ahn, M. Baumgardner, and L. Biehl, “Delineation of soil vari- ability using geostatistics and fuzzy clustering analyses of hyperspectral data,”Soil Science Society of America Journal, vol. 63, no. 1, pp. 142– 150, 1999 on Pith
  19. [19] G. Zhou, F. Yang, and J. Xiao, “Study on pixel entanglement theory for imagery classification,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–18, 2022 on Pith
  20. [20] H. B ¨acklund, A. Hedblom, and N. Neijman, “A density-based spatial clustering of application with noise,”Data Mining TNM033, vol. 33, pp. 11–30, 2011 on Pith
show 29 more
  1. [21] A. Rodriguez and A. Laio, “Clustering by fast search and find of density peaks,”Science, vol. 344, no. 6191, pp. 1492–1496, 2014 on Pith
  2. [22] R. Wang, F. Nie, and W. Yu, “Fast spectral clustering with anchor graph for large hyperspectral images,”IEEE Geoscience and Remote Sensing Letters, vol. 14, no. 11, pp. 2003–2007, 2017 on Pith
  3. [23] E. Elhamifar and R. Vidal, “Sparse subspace clustering: Algorithm, theory, and applications,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 11, pp. 2765–2781, 2013 on Pith
  4. [24] C. You, C.-G. Li, D. P. Robinson, and R. Vidal, “Oracle based active set algorithm for scalable elastic net subspace clustering,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 3928–3937 on Pith
  5. [25] K. Rafiezadeh Shahi, M. Khodadadzadeh, L. Tusa, P. Ghamisi, R. Tolosana-Delgado, and R. Gloaguen, “Hierarchical sparse subspace clustering (HESSC): An automatic approach for hyperspectral image analysis,”Remote Sensing, vol. 12, no. 15, p. 2421, 2020 on Pith
  6. [26] H. Zhao, F. Zhou, L. Bruzzone, R. Guan, and C. Yang, “Superpixel-level global and local similarity graph-based clustering for large hyperspectral images,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–16, 2021 on Pith
  7. [27] Y . Cai, Z. Zhang, P. Ghamisi, Y . Ding, X. Liu, Z. Cai, and R. Gloaguen, “Superpixel contracted neighborhood contrastive subspace clustering network for hyperspectral images,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–13, 2022 on Pith
  8. [28] Y . Zhao, Z. Bi, P. Zhu, A. Yuan, and X. Li, “Deep spectral clustering with projected adaptive feature selection,”IEEE Transactions on Geoscience and Remote Sensing, vol. 63, pp. 1–12, 2025 on Pith
  9. [29] A. Ghosh,Data Science and Cases in Sustainability: Pattern Recognition and Machine Learning. Springer Nature Singapore, 2026. [Online]. Available: https://books.google.co.in/books?id= Ghf0QEACAAJ on Pith
  10. [30] H. Zhai, H. Zhang, P. Li, and L. Zhang, “Hyperspectral image clustering: Current achievements and future lines,”IEEE Geoscience and Remote Sensing Magazine, vol. 9, no. 4, pp. 35–67, 2021 on Pith
  11. [31] D. Chakraborty, V . Narayanan, and A. Ghosh, “Integration of deep feature extraction and ensemble learning for outlier detection,”Pattern Recognition, vol. 89, pp. 161–171, 2019 on Pith
  12. [32] C.-C. Hsu and C.-W. Lin, “Cnn-based joint clustering and representation learning with feature drift compensation for large-scale image data,” IEEE Transactions on Multimedia, vol. 20, no. 2, pp. 421–429, 2017 on Pith
  13. [33] Y . Zhao and X. Li, “Deep spectral clustering with regularized linear embedding for hyperspectral image clustering,”IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–11, 2023 on Pith
  14. [34] X. Chen, Y . Zhang, X. Feng, X. Jiang, and Z. Cai, “Spectral-spatial superpixel anchor graph-based clustering for hyperspectral imagery,” IEEE Geoscience and Remote Sensing Letters, vol. 20, pp. 1–5, 2023 on Pith
  15. [35] N. P. Aghayengejeh, M. A. Balafar, J. Tanha, N. N. Khasmakhi, and S. Minaee, “Multiscale contrastive learning for node clustering based on variational graph auto-encoder,”IEEE Transactions on Neural Networks and Learning Systems, vol. 37, no. 2, pp. 646–660, Feb 2026 on Pith
  16. [36] X. Liu, T. Wang, and X. Wang, “Joint spectral–spatial representation learning for unsupervised hyperspectral image clustering,”Applied Sci- ences, vol. 15, no. 16, p. 8935, 2025. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2026 12 on Pith
  17. [37] R. Parhi and R. D. Nowak, “The role of neural network activation functions,”IEEE Signal Processing Letters, vol. 27, pp. 1779–1783, 2020 on Pith
  18. [38] M. Gra ˜na, M. A. Veganzons, and B. Ayerdi, “Hyperspectral remote sensing scenes,” https://www.ehu.eus/ccwintco/index.php/ Hyperspectral Remote Sensing Scenes, 2017, grupo de Inteligencia Computacional (GIC), University of the Basque Country on Pith
  19. [39] Y . Zhong, X. Hu, C. Luo, X. Wang, J. Zhao, and L. Zhang, “Whu-hi: Uav-borne hyperspectral with high spatial resolution (h2) benchmark datasets and classifier for precise crop identification based on deep convolutional neural network with crf,”Remote Sensing of Environment, vo... on Pith
  20. [40] M. Zhao, X. Jia, L. Fan, Y . Liang, and D.-M. Yan, “Robust ellipse fitting using hierarchical gaussian mixture models,”IEEE Transactions on Image Processing, vol. 30, pp. 3828–3843, 2021 on Pith
  21. [41] X. Yang, G. Lin, Y . Liu, F. Nie, and L. Lin, “Fast spectral embedded clustering based on structured graph learning for large-scale hyperspec- tral image,”IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1–5, 2022 on Pith
  22. [42] R. Liu, Z. Li, J. Yang, J. Sun, and Q. Liu, “Hyper-lkcnet: Exploring the utilization of large kernel convolution for hyperspectral image classifi- cation,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2025 on Pith
  23. [43] H. Zhang, X. Han, J. Deng, and W. Sun, “How to evaluate and remove the weakened bands in hyperspectral image classification,”IEEE Transactions on Geoscience and Remote Sensing, vol. 63, pp. 1–15, 2025 on Pith
  24. [44] D. Al-Alimi, Z. Cai, and M. A. Al-qaness, “Fhic: Fast hyperspectral image classification model using etr dimensionality reduction and elu activation function,”IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–17, 2023 on Pith
  25. [45] A. Datta, S. Ghosh, and A. Ghosh, “Combination of clustering and ranking techniques for unsupervised band selection of hyperspectral images,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 8, no. 6, pp. 2814–2823, 2015 on Pith
  26. [46] A. C. Thompson,Minkowski Geometry. Cambridge University Press, 1996 on Pith
  27. [47] X. Li, C. Wang, X. Zhang, and W. Sun, “Generic sao similarity measure via extended sørensen-dice index,”IEEE Access, vol. 8, pp. 66 538– 66 552, 2020 on Pith
  28. [48] R. Gonzalez and R. Woods,Digital Image Processing. Prentice Hall, 2008 on Pith
  29. [49] M. D ´ecombas, F. Dufaux, E. Renan, B. Pesquet-Popescu, and F. Cap- man, “A new object based quality metric based on sift and ssim,” in 2012 19th IEEE International Conference on Image Processing, 2012, pp. 1493–1496. BIOGRAPHY Soumi Sarkaris a research scholar at the Machine ... on Pith

Challenges and discussion

Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.

No comments yet.

Sign in to comment