Pith. sign in

REVIEW 4 major objections 5 minor 55 references

Transformer-Driven Active Transfer Learning for Cross-Hyperspectral Image Classification

T0 review · 4 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 A spatial-spectral transformer combined with active transfer learning claims gains of +2.5% to +3.8% overall accuracy over CNN, transformer, and state-space baselines on six hyperspectral benchmarks.

desk verdict The cross-dataset claim is the paper's core selling point, but the fine-tuning protocol never explains how models trained on different class sets are adapted, making the main result unreproducible as written. read the letter →

arxiv 2411.18115 v2 pith:TMNQA6I3 submitted 2024-11-27 cs.CV

Muhammad Ahmad , Francesco Mauro , Manuel Mazzara , Salvatore Distefano , Adil Mehmood Khan , Silvia Liberata Ullo This is my paper · ORCID
classification cs.CV
keywords hyperspectralimageclassificationactivelearningtransferspatial-spectraltransformercross-domainadaptationlabelefficiencydynamiclayerfreezinguncertainty-diversitysampling
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

This paper is trying to establish that a single framework can make hyperspectral image classification work in the label-scarce, cross-scene setting where most deployments actually happen. The authors propose SST-ATL, which wraps a spatial-spectral transformer in an active transfer-learning loop: a hybrid query criterion selects pixels that are both uncertain and spectrally diverse for annotation, a dynamic freezing rule adapts only the layers most affected by domain shift during fine-tuning, and self-calibrated attention reweights spectral-spatial tokens by uncertainty. They report that the method outperforms CNN-, transformer-, and state-space-model baselines by +2.5% to +3.8% overall accuracy on six benchmarks, and that fine-tuning a model trained on one dataset with only 10% of another dataset's samples keeps kappa at or above roughly 92%. If these numbers hold, the framework would make hyperspectral classification substantially cheaper in practice, since labeling is the dominant cost and scenes differ across sensors, dates, and regions.

What carries the argument

The load-bearing object is the SST-ATL pipeline: a spatial-spectral transformer encoder (3D patch embedding, sinusoidal position coding, multi-head self-attention, MLP classification head) surrounded by an active transfer loop. The hybrid query function scores each pool sample by uncertainty, $U(x_i)=-\max p(y|x_i)$, and by diversity, the mean pairwise Euclidean distance between neighboring spectral vectors, then queries the highest-scoring set. Dynamic layer freezing estimates domain shift with maximum mean discrepancy of intermediate features and freezes low-variance layers during fine-tuning, reducing compute while keeping transferable representations. Self-calibrated attention rescales attention weights by token entropy, $\mathrm{Softmax}(QK^T/\sqrt{d_k})\cdot(1+\lambda U)$, biasing the encoder toward ambiguous spatial-spectral tokens. These three mechanisms are what the ablations remove one at a time to show that each contributes to the final accuracy.

What would settle it

Run the released code with the reported 1%/49%/50% split and the reported hyperparameters, but replace the dynamic-freezing rule with a fixed rule (freeze all layers except the last) and reinitialize the classifier head for each target dataset's class set; if the +2.5% to +3.8% OA margin over baselines does not survive or the cross-dataset kappa drops below the reported range, then the unstated freezing and label-mapping details, not the framework as described, are carrying the result.

Watch

Extended reading notes

Core claim

The central claim is that active transfer learning around an SST backbone solves both halves of the cross-domain problem at once: active learning supplies the most informative target-domain labels, and transfer learning with selective adaptation supplies the source-domain prior. On the UH, PU, and SA benchmarks the paper reports ATL-SST reaching OA of 97.57%, 98.87%, and 99.48%, respectively, ahead of every compared CNN, transformer, and state-space baseline; the same-dataset learning curves show smooth gains as queried samples grow, and the ablation study attributes the largest single contribution to the hybrid uncertainty-diversity query, with secondary contributions from diversity sampling, dynamic freezing, and self-calibrated attention. In the cross-dataset experiments, an SST actively trained on one dataset and fine-tuned with 10% of another's samples keeps kappa at or above about 92% across the tested source-target pairs.

Load-bearing premise

The cross-dataset claim rests on an unspecified transfer protocol: the six datasets have different class label sets, and the paper never states how the output layer is replaced or how the dynamic-freezing rule decides which layers to freeze, so the reported margins could depend on unstated implementation choices.

Editorial extensions

If this is right

  • A user can start from one actively trained encoder and adapt it to a new hyperspectral scene with only 10% of the scene's labels, keeping $\kappa$ at or above roughly 92% on the tested pairs.
  • At the smallest label budgets tested (250 samples), the hybrid query beats random sampling by more than 3% OA on PU, UH, and SA, so the gain is largest where labeling is most expensive.
  • Ablations attribute the largest accuracy loss to removing active learning (about 2.7 OA points), with smaller but consistent drops from removing diversity sampling, dynamic freezing, or self-calibrated attention.
  • The model stays near 836K parameters and about 110K FLOPs with roughly stable training times as the labeled set grows, so repeated active-learning rounds do not require proportionally more compute.

Reading between the lines

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

  • Because the datasets carry different class label sets (for example Salinas has 16 classes while Pavia University has 9), the cross-dataset tables must involve an unstated output-layer replacement or label-semantics mapping; that choice could plausibly account for several OA points and should be reported before the transfer gains are taken at face value.
  • The dynamic-freezing rule is specified only as freezing layers whose MMD variance is low, with no threshold or schedule; until that is pinned down, the ablation 'w/o dynamic freezing' cannot be cleanly separated from ordinary fine-tuning.
  • The query strategy's early-round advantage suggests the same uncertainty-diversity criterion could transfer to few-shot or self-supervised HSI pipelines, an extension the authors mention only as possible future work.
  • If MMD-based freezing really identifies domain-invariant layers, then computing MMD on unlabeled target features could serve as a cheap domain-shift probe before any labels are queried; the paper reports no such diagnostic use.
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

4 major / 5 minor

Summary. The paper proposes SST-ATL, an active transfer learning framework for cross-dataset hyperspectral image classification. The method combines a spatial-spectral transformer backbone with an uncertainty-diversity active learning query strategy, dynamic layer freezing based on MMD, and an uncertainty-calibrated attention mechanism. Experiments are reported on six HSI datasets, including same-dataset active learning curves, cross-dataset fine-tuning tables, comparisons with CNN/Transformer/state-space baselines, and ablations. The manuscript claims consistent gains of +2.5% to +3.8% OA over baselines under constrained annotation budgets.

Significance. If the method is fully specified and reproducible, the result would be practically valuable: label-efficient cross-domain HSI classification is an important problem, and the paper evaluates on multiple standard benchmarks with a public code link. The ablation study and query-strategy comparison are useful steps toward understanding the contribution. However, as written, several load-bearing components are not specified precisely enough to reimplement, and the cross-dataset protocol that supports the main claim is not described at the architectural level. The central empirical claim is therefore not verifiable from the manuscript alone.

major comments (4)
  1. [Section VI, Tables VIII-X, Eq. (19)] The cross-dataset fine-tuning protocol is underspecified with respect to label sets. The source and target datasets have different numbers of classes (SA 16, PU 9, UH 15, HC 16, HH 22, QUH 24), yet the manuscript never states how the output layer Ofinal is replaced or re-initialized, how source and target class indices are aligned, or which C is used in the cross-entropy loss after transfer. Without this information, the reported OA/AA/kappa values in Tables VIII-X cannot be produced by the described architecture, and the paper's headline cross-dataset generalization claim is not supported.
  2. [Section III-C, Table XIV] Dynamic layer freezing is described only at a high level: 'layers with low MMD variance across domains are frozen.' No MMD threshold, layer-selection criterion, update schedule, or algorithm is given. Since the ablation in Table XIV includes a 'w/o Dynamic Freezing' variant, the implementation must exist, but as written this central component cannot be reimplemented or independently checked.
  3. [Section III-A, Eq. (9)] The calibrated attention update Acal = Softmax(QK^T/sqrt(d_k)) * (1 + lambda * U) is not well defined. U is introduced in Eq. (13) as the negative maximum predictive probability for a sample, not as a per-token vector, and the paper does not state how U is broadcast against the attention matrix or whether the result is renormalized. Multiplying a row-stochastic attention matrix by a position-dependent factor generally breaks the row-stochastic property, so the mechanism as written is underspecified.
  4. [Section III-B and Section IV] The active-query definition is internally inconsistent. Eq. (14) defines the chosen set as the argmax of Diversity(S) over all subsets of size query size, while Eq. (17) selects the top query-size entries of an argsort over per-pixel diversity; these are different selection rules unless additional structure is assumed. In addition, d_{jk} in Eq. (15) is undefined, Eq. (16) duplicates Eq. (15) with new notation, and the query percentage of 0.02% in Section IV is not reconciled with the AL rounds of 250-1000 labeled samples reported in Table XV.
minor comments (5)
  1. [Title and Section III-A] The backbone is called 'Spectral-Spectral Transformer' in Section III-A but 'Spatial-Spectral Transformer' in the abstract and elsewhere; please make the terminology consistent.
  2. [Tables XI-XIII] The header 'Tr Va Te' in Tables XI-XIII is not defined anywhere, and the per-class sample counts do not explain how the comparison methods were trained (e.g., random sampling versus active learning). Several rows report 100% OA, which makes the comparison hard to interpret without information about test-set sizes and variance.
  3. [Section V, Tables II-VII] The reported FLOPs and parameter counts vary slightly across datasets despite the same backbone architecture, and no unit or calculation formula is provided; please clarify why these values differ.
  4. [Eq. (13)] U(x_i) = -max p(y|x_i) is a negative confidence score, not an entropy; the text calls it 'entropy-based uncertainty,' which is inaccurate.
  5. [Throughout] The names 'SST-ATL' and 'ATL-SST' are used inconsistently, and Table XI contains the typo 'WaveMamab' for WaveMamba; please unify the nomenclature.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's claims rest on benchmark experiments, and no prediction reduces by construction to a fitted input or to a load-bearing self-citation.

full rationale

The paper is an empirical systems paper. Its central claim, that the proposed SST-ATL framework outperforms baselines under constrained annotation budgets, is supported by direct experiments on benchmark datasets (Tables II-VII, XI-XIII) rather than by a derivation from assumptions that already contain the conclusion. The uncertainty measure U used in the calibrated attention (Eq. 9) is also used in the active-learning query criterion (Eq. 13), so the feedback loop is deliberate design rather than a hidden circularity: the model focuses on ambiguous tokens and also queries ambiguous samples, but the reported accuracies are measured on held-out test sets, so the numbers are not forced by construction. The dynamic freezing rule (Sec. III-C) is underspecified and hard to reproduce, and the cross-dataset fine-tuning protocol (Sec. VI) does not explain how output layers with different class counts are handled, but these are correctness and reproducibility concerns, not circularity. The paper cites several of its own prior works (e.g., [43], [47], [55]) as background and baselines, but none of these citations is used to justify a uniqueness claim, to import an unverified theorem, or to define the reported result in terms of itself. There is no fitted parameter that is later renamed as a prediction, and no equation reduces to another equation by construction. Accordingly, the appropriate finding is no significant circularity.

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

The central claims depend on several hand-chosen parameters (λ, MMD threshold, query percentage, neighborhood size, data splits) and on the unstated assumption that cross-dataset fine-tuning can ignore differing class label sets. No new physical entities are introduced.

free parameters (6)
  • λ (uncertainty scaling in Acal) = not reported
    Hyperparameter controlling the influence of uncertainty in self-calibrated attention, Eq. (9). No value or sensitivity analysis is given.
  • MMD freezing threshold = not reported
    Dynamic layer freezing selects layers with low MMD variance; the threshold for 'low' is unspecified.
  • Query percentage = 0.02%
    Fraction of pool queried per AL round; set in Section IV, no ablation.
  • Neighborhood size for diversity = 3 (9 vectors)
    n_neighborhood=3 in Section IV defines the diversity computation; no ablation.
  • Train/pool/test split = 1% / 49% / 50%
    Section IV; the extreme imbalance is not justified.
  • Fine-tuning sample fraction = 10%
    Used in cross-dataset experiments (Section VI); no comparison to other fractions.
assumptions (4)
  • domain assumption MMD between intermediate features measures domain shift magnitude
    Invoked in Section III-C to decide which layers to freeze.
  • domain assumption Source and target datasets in cross-domain experiments share a compatible label space
    Implied by fine-tuning a model trained on one dataset with 10% labels from another, but datasets have different class counts (e.g., SA 16 vs PU 9); never stated or handled.
  • ad hoc to paper Uncertainty-diversity hybrid querying improves over either criterion alone
    Central design choice of the AL module, supported only by the paper's own ablation.
  • standard math Standard transformer, softmax, and cross-entropy machinery
    Background for the SST backbone, Eqs. (4)-(8), (19).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transformer-Driven Active Transfer Learning for Cross-Hyperspectral Image Classification." pith.science (2026). https://pith.science/paper/TMNQA6I3

@misc{pith2026241118115,
  author       = {Pith},
  title        = {Pith review of: Transformer-Driven Active Transfer Learning for Cross-Hyperspectral Image Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TMNQA6I3}},
  note         = {Machine review of arXiv:2411.18115}
}
read the original abstract

Hyperspectral image (HSI) classification presents inherent challenges due to high spectral dimensionality, significant domain shifts, and limited availability of labeled data. To address these issues, we propose a novel Active Transfer Learning (ATL) framework built upon a Spatial-Spectral Transformer (SST) backbone. The framework integrates multistage transfer learning with an uncertainty-diversity-driven active learning mechanism that strategically selects highly informative and diverse samples for annotation, thereby significantly reducing labeling costs and mitigating sample redundancy. A dynamic layer freezing strategy is introduced to enhance transferability and computational efficiency, enabling selective adaptation of model layers based on domain shift characteristics. Furthermore, we incorporate a self-calibrated attention mechanism that dynamically refines spatial and spectral weights during adaptation, guided by uncertainty-aware feedback. A diversity-promoting sampling strategy ensures broad spectral coverage among selected samples, preventing overfitting to specific classes. Extensive experiments on benchmark cross-domain HSI datasets demonstrate that the proposed SST-ATL framework achieves superior classification performance compared to conventional approaches. The source code is publicly available at https://github.com/mahmad000/ATL-SST.

Figures

Figures reproduced from arXiv: 2411.18115 by the authors.

Figure 1
Figure 1. Overview of the proposed ATL-SST framework for HSI classification. The model first trains an SST encoder by extracting [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. PU dataset: Ground truth maps corresponding to varying numbers of training samples. model’s complexity at approximately 110,592 FLOPs and 836,816 parameters. These values confirm the model’s high computational efficiency, which is an essential consideration for large-scale applications. Finally, [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 2
Figure 2. UH dataset: Ground truth maps corresponding to varying numbers of training samples. The results presented in Table III and 3 collectively high￾light the model’s performance progression on the PU dataset as training samples increase. Training and testing times remain relatively stable across sample sizes, indicating computational efficiency; a total of 108,800 FLOPs and 835,017 parameters suggests a well-optimized mo… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: SA dataset: Ground truth maps corresponding to varying numbers of training samples. Table V provides a comprehensive overview of the HC dataset’s classification performance and computational cost. The computational metrics reveal a gradual increase in training time wit…
Figure 6
Figure 6. Figure 6: HH dataset: Ground truth maps corresponding to varying numbers of training samples. ples increases, corroborating the quantitative improvements observed in Table VII. TABLE VII: QUH Dataset: Accuracy improvements as the sample count increases per iteration. Class Numbe…
Figure 5
Figure 5. Figure 5: HC dataset: Ground truth maps corresponding to varying numbers of training samples. ple, training times range from 272.05s to 291.68s, and testing times fluctuate minimally between 22.42s and 23.42s. For instance, OA increased from 87.83% to 98.97%, while AA rose from …
Figure 8
Figure 8. Figure 8: SST trained on SA was fine-tuned and tested on PU, HC, HH, and UH datasets, respectively. across datasets, with κ values consistently above 93%, demon￾strating effective transfer learning. Testing times vary sig￾nificantly across datasets due to differences in data vol…
Figure 9
Figure 9. Figure 9: SST trained on PU was fine-tuned and tested on SA, HC, HH, and UH datasets, respectively. TABLE X: The OA, AA, and κ accuracies for cross-dataset classification were obtained using the actively trained SST on the UH dataset and evaluated across other datasets. Metric S…
Figure 10
Figure 10. Figure 10: SST trained on UH was fine-tuned and tested on SA, HC, HH, and PU datasets, respectively. other datasets, the model demonstrates strong transferability, achieving an OA of 98.47% on SA and 97.75% on PU. Although the performance is slightly reduced for HC and HH, the O…
Figure 11
Figure 11. Figure 11: UH dataset: Ground truth maps for comparative methods alongside the proposed ATL-SST. TABLE XII: PU dataset: Comparison of various HSI classification methods highlighting performance metrics. Samples State-of-the-art Comparative Methods ATL-SST Tr Va Te AGCN [51] HViT…
Figure 12
Figure 12. Figure 12: PU dataset: Ground truth maps for comparative methods alongside the proposed ATL-SST. TABLE XIII: SA Comparison of various HSI classification methods highlighting performance metrics. Samples State-of-the-art Comparative Methods ATL-SST Tr Va Te AGCN [51] HViT [52] SS…
Figure 13
Figure 13. Figure 13: SA dataset: Ground truth maps for comparative methods alongside the proposed ATL-SST. models like HybViT and SST, while being significantly faster than WaveMamba and PyFormer, which require 1734.50s and 1469.81s, respectively. Qualitative results in [PITH_FULL_IMAGE:…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 45 canonical work pages

  1. [1]

    Hyperspectral remote sensing data analysis and future challenges,

    J. M. Bioucas-Dias, A. Plaza, G. Camps-Valls, P. Scheunders, N. Nasrabadi, and J. Chanussot, “Hyperspectral remote sensing data analysis and future challenges,” IEEE Geoscience and Remote Sensing Magazine, vol. 1, no. 2, pp. 6–36, 2013

    2013

  2. [2]

    Graph-infused hybrid vision transformer: Advancing geoai for enhanced land cover classification,

    M. H. F. Butt, J. P. Li, M. Ahmad, and M. A. F. Butt, “Graph-infused hybrid vision transformer: Advancing geoai for enhanced land cover classification,” International Journal of Applied Earth Observation and Geoinformation , vol. 129, p. 103773, 2024. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S1569843224001274

    work page 2024

  3. [3]

    High-resolution hyperspectral imaging using low-cost components: Application within environmental monitoring scenarios,

    M. B. Stuart, M. Davies, M. J. Hobbs, T. D. Pering, A. J. McGonigle, and J. R. Willmott, “High-resolution hyperspectral imaging using low-cost components: Application within environmental monitoring scenarios,” Sensors, vol. 22, no. 12, p. 4652, 2022

    work page 2022

  4. [4]

    Prediction of microbial spoilage and shelf-life of bakery products through hyperspectral imaging,

    Z. Saleem, M. H. Khan, M. Ahmad, A. Sohaib, H. Ayaz, and M. Maz- zara, “Prediction of microbial spoilage and shelf-life of bakery products through hyperspectral imaging,” IEEE Access , vol. 8, pp. 176 986– 176 996, 2020

    work page 2020

  5. [5]

    Hyperspectral imaging-based unsupervised adulterated red chili content transformation for classification: Identification of red chili adulterants,

    M. H. Khan, Z. Saleem, M. Ahmad, A. Sohaib, H. Ayaz, M. Mazzara, and R. A. Raza, “Hyperspectral imaging-based unsupervised adulterated red chili content transformation for classification: Identification of red chili adulterants,” Neural Computing and Applications , vol. 33, no. 21, pp. 14 507–14 521, 2021

    work page 2021

  6. [6]

    Hyperspectral imaging for minced meat classification using nonlinear deep features,

    H. Ayaz, M. Ahmad, M. Mazzara, and A. Sohaib, “Hyperspectral imaging for minced meat classification using nonlinear deep features,” Applied Sciences, vol. 10, no. 21, p. 7783, 2020. JOURNAL OF LATEX CLASS FILES 13

    work page 2020

  7. [7]

    A fast and compact hybrid cnn for hyperspectral imaging-based bloodstain classification,

    M. H. F. Butt, H. Ayaz, M. Ahmad, J. P. Li, and R. Kuleev, “A fast and compact hybrid cnn for hyperspectral imaging-based bloodstain classification,” in 2022 IEEE Congress on Evolutionary Computation (CEC), 2022, pp. 1–8

    work page 2022

  8. [8]

    Hyperspectral imaging for bloodstain identification,

    M. Zulfiqar, M. Ahmad, A. Sohaib, M. Mazzara, and S. Distefano, “Hyperspectral imaging for bloodstain identification,” Sensors, vol. 21, no. 9, p. 3045, 2021

    work page 2021

Show all 55 references
  1. [9]

    Intelligent tumor tissue classification for hybrid health care units,

    M. H. F. Butt, J. P. Li, J. C. Ji, W. Riaz, N. Anwar, F. F. Butt, M. Ahmad, A. Saboor, A. Ali, and M. Y . Uddin, “Intelligent tumor tissue classification for hybrid health care units,” Frontiers in Medicine, vol. 11, 2024. [Online]. Available: https://www.frontiersin.org/journ...

    2024

  2. [10]

    A review on hyperspectral imagery application for lithological mapping and mineral prospecting: Machine learning techniques and future prospects,

    S. Hajaj, A. El Harti, A. B. Pour, A. Jellouli, Z. Adiri, and M. Hashim, “A review on hyperspectral imagery application for lithological mapping and mineral prospecting: Machine learning techniques and future prospects,” Remote Sensing Applications: Society and Environment , v...

    2024

  3. [11]

    Deep learning techniques for hyperspectral image analysis in agriculture: A review,

    M. F. Guerri, C. Distante, P. Spagnolo, F. Bougourzi, and A. Taleb- Ahmed, “Deep learning techniques for hyperspectral image analysis in agriculture: A review,” ISPRS Open Journal of Photogrammetry and Remote Sensing , vol. 12, p. 100062, 2024. [Online]. Available: https://www...

    2024

  4. [12]

    Pyramid hierarchical spatial-spectral transformer for hyperspectral image classification,

    M. Ahmad, M. H. F. Butt, M. Mazzara, S. Distefano, A. M. Khan, and H. A. Altuwaijri, “Pyramid hierarchical spatial-spectral transformer for hyperspectral image classification,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 17, pp. 17 6...

    2024

  5. [13]

    Hyperspectral image classification using groupwise separable convolutional vision transformer network,

    Z. Zhao, X. Xu, S. Li, and A. Plaza, “Hyperspectral image classification using groupwise separable convolutional vision transformer network,” IEEE Transactions on Geoscience and Remote Sensing , vol. 62, pp. 1– 17, 2024

    2024

  6. [14]

    Multi-head spatial-spectral mamba for hyperspectral image classification,

    M. Ahmad, M. H. F. Butt, M. Usama, H. A. Altuwaijri, M. Mazzara, S. Distefano, and A. M. Khan, “Multi-head spatial-spectral mamba for hyperspectral image classification,” Remote Sensing Letters , vol. 16, no. 4, pp. 339–353, 2025. [Online]. Available: https: //doi.org/10.1080/...

    2025

  7. [15]

    Bridging cnn and transformer with cross-attention fusion network for hyperspectral image classification,

    F. Xu, S. Mei, G. Zhang, N. Wang, and Q. Du, “Bridging cnn and transformer with cross-attention fusion network for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 62, pp. 1–14, 2024

    2024

  8. [16]

    Differ- ential attention with enhanced squeeze-and-excitation for hyperspectral image classification,

    S. Sohail, M. Usama, U. Ghous, M. Mazzara, and M. Ahmad, “Differ- ential attention with enhanced squeeze-and-excitation for hyperspectral image classification,” IEEE Geoscience and Remote Sensing Letters , vol. 22, pp. 1–5, 2025

    2025

  9. [17]

    Graphgst: Graph generative structure-aware transformer for hyperspec- tral image classification,

    M. Jiang, Y . Su, L. Gao, A. Plaza, X.-L. Zhao, X. Sun, and G. Liu, “Graphgst: Graph generative structure-aware transformer for hyperspec- tral image classification,” IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–16, 2024

    2024

  10. [18]

    Scformer: Spectral coordinate transformer for cross-domain few-shot hyperspectral image classification,

    J. Li, Z. Zhang, R. Song, Y . Li, and Q. Du, “Scformer: Spectral coordinate transformer for cross-domain few-shot hyperspectral image classification,” IEEE Transactions on Image Processing , vol. 33, pp. 840–855, 2024

    2024

  11. [19]

    Spectralformer: Rethinking hyperspectral image classification with transformers,

    D. Hong, Z. Han, J. Yao, L. Gao, B. Zhang, A. Plaza, and J. Chanus- sot, “Spectralformer: Rethinking hyperspectral image classification with transformers,” IEEE Transactions on Geoscience and Remote Sensing , vol. 60, pp. 1–15, 2022

    2022

  12. [20]

    Tccu-net: Transformer and cnn collaborative unmixing network for hyperspectral image,

    J. Chen, C. Yang, L. Zhang, L. Yang, L. Bian, Z. Luo, and J. Wang, “Tccu-net: Transformer and cnn collaborative unmixing network for hyperspectral image,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 17, pp. 8073–8089, 2024

    2024

  13. [22]

    A dual-branch multiscale transformer network for hyperspectral image classification,

    C. Shi, S. Yue, and L. Wang, “A dual-branch multiscale transformer network for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 62, pp. 1–20, 2024

    2024

  14. [23]

    Sparse self- attention transformer for image inpainting,

    W. Huang, Y . Deng, S. Hui, Y . Wu, S. Zhou, and J. Wang, “Sparse self- attention transformer for image inpainting,” Pattern Recognition, vol. 145, p. 109897, 2024

    2024

  15. [24]

    Image fusion for the novelty rotating synthetic aperture system based on vision transformer,

    Y . Sun, X. Zhi, S. Jiang, G. Fan, X. Yan, and W. Zhang, “Image fusion for the novelty rotating synthetic aperture system based on vision transformer,” Information Fusion, vol. 104, p. 102163, 2024

    2024

  16. [25]

    Spatial–spectral transformer with cross-attention for hyperspectral image classification,

    Y . Peng, Y . Zhang, B. Tu, Q. Li, and W. Li, “Spatial–spectral transformer with cross-attention for hyperspectral image classification,” IEEE Trans- actions on Geoscience and Remote Sensing , vol. 60, pp. 1–15, 2022

    2022

  17. [26]

    A survey: Deep learning for hyperspectral image classification with few labeled samples,

    S. Jia, S. Jiang, Z. Lin, N. Li, M. Xu, and S. Yu, “A survey: Deep learning for hyperspectral image classification with few labeled samples,” Neurocomputing, vol. 448, pp. 179–204, 2021. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0925231221004033

    2021

  18. [27]

    A survey of deep active learning,

    P. Ren, Y . Xiao, X. Chang, P.-Y . Huang, Z. Li, B. B. Gupta, X. Chen, and X. Wang, “A survey of deep active learning,” ACM Comput. Surv. , vol. 54, no. 9, Oct. 2021. [Online]. Available: https://doi.org/10.1145/3472291

    2021 doi

  19. [28]

    Spatial- prior generalized fuzziness extreme learning machine autoencoder-based active learning for hyperspectral image classification,

    M. Ahmad, S. Shabbir, D. Oliva, M. Mazzara, and S. Distefano, “Spatial- prior generalized fuzziness extreme learning machine autoencoder-based active learning for hyperspectral image classification,” Optik, vol. 206, p. 163712, 2020. [Online]. Available: https://www.sciencedir...

    2020

  20. [29]

    Dbral: A novel uncertainty-based active learning based on deep-broad learning for medical image classification,

    H. Wu, Y . Zhong, G. Han, J. Lin, Z. Liu, and C. Han, “Dbral: A novel uncertainty-based active learning based on deep-broad learning for medical image classification,” in Artificial Neural Networks and Machine Learning – ICANN 2024: 33rd International Conference on Artificial ...

    2024

  21. [30]

    Active learning for hyperspectral image classification: A comparative review,

    R. Thoreau, V . Achard, L. Risser, B. Berthelot, and X. Briottet, “Active learning for hyperspectral image classification: A comparative review,” IEEE Geoscience and Remote Sensing Magazine , vol. 10, no. 3, pp. 256–278, 2022

    2022

  22. [31]

    Hyperspectral image classification with fuzzy spatial-spectral class discriminate infor- mation,

    M. Ahmad, M. Usama, S. Distefano, and M. Mazzara, “Hyperspectral image classification with fuzzy spatial-spectral class discriminate infor- mation,” in 2024 IEEE International Conference on Image Processing (ICIP), 2024, pp. 2285–2291

    2024

  23. [32]

    An end-to-end active learning framework for limited labelled hyperspectral image classification,

    A. C. Karaca and G. Bilgin, “An end-to-end active learning framework for limited labelled hyperspectral image classification,” International Journal of Remote Sensing , vol. 46, no. 8, pp. 3179–3206, 2025. [Online]. Available: https://doi.org/10.1080/01431161.2025.2467294

    2025

  24. [33]

    Few-shot hyperspectral image classi- fication with unknown classes using multitask deep learning,

    S. Liu, Q. Shi, and L. Zhang, “Few-shot hyperspectral image classi- fication with unknown classes using multitask deep learning,” IEEE Transactions on Geoscience and Remote Sensing , vol. 59, no. 6, pp. 5085–5102, 2021

    2021

  25. [34]

    Class-wise graph embedding- based active learning for hyperspectral image classification,

    X. Liao, B. Tu, J. Li, and A. Plaza, “Class-wise graph embedding- based active learning for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–13, 2023

    2023

  26. [35]

    Hyperspectral image classification with multi-attention transformer and adaptive superpixel segmentation-based active learning,

    C. Zhao, B. Qin, S. Feng, W. Zhu, W. Sun, W. Li, and X. Jia, “Hyperspectral image classification with multi-attention transformer and adaptive superpixel segmentation-based active learning,” IEEE Transac- tions on Image Processing , vol. 32, pp. 3606–3621, 2023

    2023

  27. [36]

    Active learning with co-auxiliary learning and multi-level diversity for image classification,

    Z. Wang, Z. Chen, and B. Du, “Active learning with co-auxiliary learning and multi-level diversity for image classification,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 33, no. 8, pp. 3899– 3911, 2023

    2023

  28. [37]

    Adversarial domain alignment with contrastive learning for hyperspectral image classifica- tion,

    F. Liu, W. Gao, J. Liu, X. Tang, and L. Xiao, “Adversarial domain alignment with contrastive learning for hyperspectral image classifica- tion,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–20, 2023

    2023

  29. [38]

    Deep learning segmentation of seismic facies based on proximity constraint strategy: Innovative application of uma-net model,

    F. Deng, R. Liang, W. Luo, and G. Zhang, “Deep learning segmentation of seismic facies based on proximity constraint strategy: Innovative application of uma-net model,” IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–17, 2024

    2024

  30. [39]

    Unsupervised blind spectral–spatial cross-super-resolution network for hsi and msi fusion,

    H. wu, S. Wu, K. Zhang, X. Liu, S. Shi, and C. Bian, “Unsupervised blind spectral–spatial cross-super-resolution network for hsi and msi fusion,” IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–14, 2024

    2024

  31. [40]

    A spectral–spatial fusion transformer network for hyperspectral image classification,

    D. Liao, C. Shi, and L. Wang, “A spectral–spatial fusion transformer network for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–16, 2023

    2023

  32. [41]

    Els2t: Efficient lightweight spectral–spatial transformer for hyperspectral image classifi- cation,

    S. Zhang, J. Zhang, X. Wang, J. Wang, and Z. Wu, “Els2t: Efficient lightweight spectral–spatial transformer for hyperspectral image classifi- cation,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–16, 2023

    2023

  33. [42]

    A multi-range spectral-spatial transformer for hyperspectral image classification,

    L. Zhang, Y . Wang, L. Yang, J. Chen, Z. Liu, J. Wang, L. Bian, and C. Yang, “A multi-range spectral-spatial transformer for hyperspectral image classification,” Infrared Physics & Technology , vol. 135, p. 104983, 2023. [Online]. Available: https://www.sciencedirect.com/ scie...

    2023

  34. [43]

    Waveformer: Spec- tral–spatial wavelet transformer for hyperspectral image classification,

    M. Ahmad, U. Ghous, M. Usama, and M. Mazzara, “Waveformer: Spec- tral–spatial wavelet transformer for hyperspectral image classification,” IEEE Geoscience and Remote Sensing Letters , vol. 21, pp. 1–5, 2024

    2024

  35. [44]

    Active learning-driven siamese network for hyperspectral image classification,

    X. Di, Z. Xue, and M. Zhang, “Active learning-driven siamese network for hyperspectral image classification,” Remote Sensing, vol. 15, no. 3,

  36. [45]

    Active transfer learning network: A unified deep joint spectral–spatial feature learning model for hyperspectral image classification,

    C. Deng, Y . Xue, X. Liu, C. Li, and D. Tao, “Active transfer learning network: A unified deep joint spectral–spatial feature learning model for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 57, no. 3, pp. 1741–1754, 2019

    2019

  37. [46]

    Active-learning- incorporated deep transfer learning for hyperspectral image classifica- tion,

    J. Lin, L. Zhao, S. Li, R. Ward, and Z. J. Wang, “Active-learning- incorporated deep transfer learning for hyperspectral image classifica- tion,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 11, no. 11, pp. 4048–4062, 2018

    2018

  38. [47]

    A disjoint samples-based 3d-cnn with active transfer learning for hyperspectral image classification,

    M. Ahmad, U. Ghous, D. Hong, A. M. Khan, J. Yao, S. Wang, and J. Chanussot, “A disjoint samples-based 3d-cnn with active transfer learning for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 60, pp. 1–16, 2022

    2022

  39. [48]

    Active learning for visual image classification method based on transfer learning,

    J. Yang, S. Li, and W. Xu, “Active learning for visual image classification method based on transfer learning,” IEEE Access, vol. 6, pp. 187–198, 2018

    2018

  40. [49]

    Hyperspectral image classification with convolutional neural network and active learning,

    X. Cao, J. Yao, Z. Xu, and D. Meng, “Hyperspectral image classification with convolutional neural network and active learning,” IEEE Transac- tions on Geoscience and Remote Sensing , vol. 58, no. 7, pp. 4604–4616, 2020

    2020

  41. [50]

    Heterogeneous transfer learning for hyperspectral image classification based on convolutional neural net- work,

    X. He, Y . Chen, and P. Ghamisi, “Heterogeneous transfer learning for hyperspectral image classification based on convolutional neural net- work,” IEEE Transactions on Geoscience and Remote Sensing , vol. 58, no. 5, pp. 3246–3263, 2020

    2020

  42. [51]

    Attention graph convolutional network for disjoint hyperspectral image classification,

    A. Jamali, S. K. Roy, D. Hong, P. M. Atkinson, and P. Ghamisi, “Attention graph convolutional network for disjoint hyperspectral image classification,” IEEE Geoscience and Remote Sensing Letters , vol. 21, pp. 1–5, 2024

    2024

  43. [52]

    Hybrid multi- scale spatial–spectral transformer for hyperspectral image classification,

    Y . He, B. Tu, B. Liu, Y . Chen, J. Li, and A. Plaza, “Hybrid multi- scale spatial–spectral transformer for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 62, pp. 1– 18, 2024

    2024

  44. [53]

    Centerformer: A center spatial–spectral attention transformer network for hyperspectral image classification,

    C. Jia, X. Zhang, H. Meng, S. Xia, and L. Jiao, “Centerformer: A center spatial–spectral attention transformer network for hyperspectral image classification,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 18, pp. 5523–5539, 2025

    2025

  45. [54]

    S2mamba: A spa- tial–spectral state space model for hyperspectral image classification,

    G. Wang, X. Zhang, Z. Peng, T. Zhang, and L. Jiao, “S2mamba: A spa- tial–spectral state space model for hyperspectral image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 63, pp. 1– 13, 2025

    2025

  46. [55]

    Wavemamba: Spatial-spectral wavelet mamba for hyperspectral image classification,

    M. Ahmad, M. Usama, M. Mazzara, and S. Distefano, “Wavemamba: Spatial-spectral wavelet mamba for hyperspectral image classification,” IEEE Geoscience and Remote Sensing Letters , pp. 1–1, 2024

    2024

  47. [2023]

    Available: https://www.mdpi.com/2072-4292/15/3/752

    [Online]. Available: https://www.mdpi.com/2072-4292/15/3/752

Pith tools

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