Pith. sign in

REVIEW 4 major objections 5 minor 63 references

ECP-Mamba: An Efficient Multi-scale Self-supervised Contrastive Learning Method with State Space Model for PolSAR Image Classification

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Radar pixel map hits 99.70% with a spiral-scan Mamba

desk verdict Useful Mamba-for-PolSAR adaptation with a sensible spiral scan and a clean multi-scale pretext task, but the headline SOTA numbers are not yet trustworthy because training and test pixels overlap with no error bars. read the letter →

arxiv 2506.01040 v1 pith:6H54NVEM submitted 2025-06-01 cs.CV

Zuzheng Kuang , Haixia Bi , Chen Xu , Jian Sun This is my paper
classification cs.CV
keywords PolSARimageclassificationself-supervisedlearningcontrastivestatespacemodelMambaspiralscanstrategymulti-scalefeaturefusionself-distillation
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

ECP-Mamba claims to classify polarimetric SAR (PolSAR) imagery pixel by pixel with state-of-the-art accuracy while needing very few labels and much less computation than Transformer-based models. The method rests on two ideas: a self-supervised pre-training task that aligns local and global views of the same pixel through self-distillation, without negative samples, and a Mamba backbone whose spiral scan orders tokens so that context near the central pixel accumulates at the end of the sequence, right before a learnable class token. A lightweight Cross Mamba module then exchanges class tokens between the local and global branches so multi-scale features reinforce each other at minimal extra cost. On the Flevoland 1989 benchmark the paper reports 99.70% overall accuracy, 99.64% average accuracy, and a Kappa coefficient of 99.62e-2 at a 0.2% label sampling rate, and reports that ECP-Mamba beats competing self-supervised methods on all four datasets studied.

What carries the argument

The load-bearing mechanism is the spiral scan strategy: a patch's tokens are reordered by a spiral traversal that starts at the top-left corner and winds inward, then a learnable tail class token is appended and sinusoidal positional embeddings are added, so the central pixel ends up adjacent to the class token where the SSM's hidden state has accumulated the most context. Around this sits the Bidirectional Polarimetric Selective Scanning block, a forward-backward selective SSM with residual connection, and the Cross Mamba module, which fuses scales by exchanging tail class tokens between branches. The self-supervised loss is a negative-free self-distillation objective: softmax cross-entropy between the local branch's sharpened distribution and the global branch's centered, sharpened distribution, with the global branch updated by exponential moving average.

What would settle it

Re-run the Flevoland 1989 experiment at a 0.2% sampling rate but evaluate each pixel only when its entire local and global patch is disjoint from every training pixel, or hold out whole contiguous regions; if overall accuracy drops well below the reported 99.70%, part of the result is carried by shared receptive fields rather than by generalization.

Watch

Extended reading notes

Core claim

The paper's central claim is that a selective state space model can be adapted to per-pixel PolSAR classification by making the scanning order respect the causal structure of the task: pixels closer to the central pixel carry more information about its label, so the spiral scan places them near the end of the sequence, immediately before a learnable tail class token that summarizes the whole patch. Classification fine-tuning fuses local 16×16 and global 32×32 views through a Cross Mamba module that swaps the tail class tokens of the two branches, and the contrastive pre-training stage learns without labels by predicting the global view's sharpened, centered softmax distribution from the local view, with only the local branch receiving gradients and the global branch updated as an exponential moving average. The paper reports that this combination achieves 99.70% OA, 99.64% AA, and 99.62e-2 Kappa on Flevoland 1989, and its ablations show that spiral scan, multi-scale branches, Cross Mamba, contrastive pre-training, and data augmentation each contribute to that result.

Load-bearing premise

The reported gains assume that randomly choosing 0.2% of labeled pixels to train on and then predicting every pixel, with local and global windows that overlap between training and test locations, does not inflate the scores; the paper does not say that training and test patches are kept spatially separate.

Editorial extensions

If this is right

  • If ECP-Mamba is right, per-pixel PolSAR classification no longer needs Transformer-scale attention; in the paper's ablations the spiral-scan Mamba reaches higher accuracy than a ViT baseline with about one tenth of the FLOPs.
  • The reported sampling-rate experiments imply self-supervised pre-training carries most of the labeling burden: at 0.05% labels on Flevoland 1989 the full method reports 97.99% overall accuracy while the Vim baseline reports 71.45%.
  • The local-to-global self-distillation objective is presented as a reusable recipe for any pixel-centric task: pre-train on unlabeled image patches by predicting a global view from a local view, with no negative pairs.
  • The tail class token position is load-bearing: performance on Flevoland 1989 peaks near the center of the spiral and drops when the token is placed far from the central pixel, so scan order should be tuned for Mamba-based per-pixel classifiers.

Reading between the lines

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

  • Beyond the paper: the spiral-scan principle should transfer to other per-pixel remote-sensing tasks, such as land-cover mapping, change detection, and hyperspectral classification, wherever the label belongs to the center of a window and near context matters more than far context.
  • Beyond the paper: a controlled comparison with a CNN or Transformer trained under the identical multi-scale self-distillation pre-text would isolate how much of the reported gain comes from the SSM and spiral scan rather than from the pre-training objective.
  • Beyond the paper: because the paper does not describe a patch-disjoint train/test split, the headline numbers should be re-checked under spatial separation; only then can the true gain over leakage-prone evaluation be quantified.
  • Beyond the paper: the negative-free self-distillation recipe can be stress-tested directly by removing the EMA and centering operations and measuring whether the global branch collapses, which would show whether the stability machinery is essential to the result.
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 ECP-Mamba, a framework for PolSAR image classification that combines multi-scale self-supervised contrastive learning with a state space model (Mamba) backbone. The method introduces a spiral scan strategy to order patch tokens by causal relevance to the central pixel, a bidirectional selective scanning block, and a Cross Mamba module for fusing local and global features. The authors report state-of-the-art results on four PolSAR benchmarks, with a headline overall accuracy of 99.70% on Flevoland 1989 using only 0.2% labeled pixels. Ablation studies and parameter analyses are provided, and the code is promised.

Significance. If the reported results are trustworthy, ECP-Mamba would be a meaningful advance for label-scarce PolSAR classification, being the first adaptation of Mamba/SSM to this task and showing large gains over prior contrastive and supervised baselines at extremely low sampling rates. The proposed spiral scan and Cross Mamba are intuitively motivated and the reported FLOPs are low. However, the central empirical claim rests on an evaluation protocol that may allow training/test leakage through overlapping receptive fields, and the results lack error bars or multiple-seed variation. The significance of the contribution is therefore currently conditional on the protocol being corrected.

major comments (4)
  1. [Section VII-A and Algorithm 2] The evaluation protocol does not prevent overlap between training and test patches. Algorithm 2 states that the classifier is applied to the complete dataset D and outputs O, while fine-tuning uses a randomly selected labeled subset D′. With local patch size k=16 and global patch size K=32 centered on each pixel, a test pixel located even one pixel away from a training pixel shares 31 of 32 columns/rows of its global patch and large parts of its local patch. If the training pixels themselves are also evaluated, or if no spatial buffer is used, the model can memorize the few hundred training patches and propagate their labels to neighboring pixels by patch similarity. The paper does not state that training pixels are excluded from evaluation, nor does it describe any patch-disjoint split or spatial buffer. This threatens the validity of the headline accuracies, including the 99.70% OA on Flevoland 1989. Please clarify whether training pixels are included in the evaluation set and report results with a spatial buffer (e.g., excluding all pixels whose receptive field overlaps a training pixel) and with training pixels removed from the test set.
  2. [Section VII-D1 and Table III] The network hyperparameters (patch sizes, hidden dimension, number of BPSS layers, etc.) are selected on the Flevoland 1989 dataset in the parameter analysis, and the same dataset is then used for the headline SOTA claim in Table III. This creates a risk of overfitting to a single test set. While the method is also evaluated on three other datasets, the default settings are not validated for sensitivity across all datasets. Please either perform parameter selection on a validation split of Flevoland 1989 or on a different dataset, and report results for the other datasets with per-dataset tuned parameters alongside the fixed-default results.
  3. [Tables III, IV, V, VI and Table VIII] All experimental numbers are given without error bars or repeated runs. At sampling rates as low as 0.2% (and 0.05% for San Francisco), the random choice of training pixels can cause large variance across seeds; for example, the Flevoland 1989 buildings class has only 2 labeled instances, so a single random draw may not be representative. The paper should report the mean and standard deviation over at least 5 independent random splits, and ideally over multiple training seeds, for both the proposed method and the main baselines. Without this, the claimed margins over PiCL (e.g., 4.25% OA on Flevoland 1989) cannot be assessed statistically.
  4. [Section IV-D versus Section VII-A] There is a direct inconsistency in the EMA momentum parameter λ: Section IV-D states that λ follows a cosine schedule increasing from 0.9995 to 1 during pre-training, while Section VII-A states 'The momentum parameter λ is 0.996.' The paper must state which schedule is actually used, since this affects reproducibility and possibly the reported results.
minor comments (5)
  1. [Section VII-D1] In the paragraph preceding Table VII, the text says 'the number of BBPS blocks' but the block is defined as BPSS in Section IV-C2. Please fix the typo.
  2. [Equation (18)] The first expression for the PolyLoss uses (1+ε1)(1−ȳ(i)) and a summation over j, while the second simplifies to −y(i)log(ȳ(i))+ε(1−ȳ(i)). The equivalence is not shown and the first form appears to contain a typo (ε1 vs ε, and the y(i) term). Please rewrite the equation and show the expansion.
  3. [Section V-B] The sentence 'Only the final tokens from w′ and W′, denoted as ct and Ct, are used' uses lowercase ct for both the class token and the token from the local branch; in Section IV-C1 ct is defined as the learnable tail class token. Consider renaming the fine-tuning tokens (e.g., c_t and C_t) to avoid ambiguity.
  4. [Section VII-A] The paper says 'The learning rate and momentum parameters undergo decay following a cosine annealing strategy [60]' but the EMA momentum λ is described as fixed at 0.996 in the same paragraph. Please clarify which momentum is decayed and which is fixed.
  5. [Table II] The FLOPs for CL + Multi-scale Efficient Mamba are listed as '-' with the note that FLOPs do not apply to the self-supervised method. This is confusing because the fine-tuning stage still uses the same architecture; if FLOPs are only reported for the fine-tuning forward pass, please state that explicitly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the SOTA claim is empirical, and the few self-citations are non-load-bearing baselines or protocol settings.

full rationale

ECP-Mamba's central claim, state-of-the-art PolSAR classification accuracy, is supported by measured results on benchmark datasets (Tables III--VI) rather than derived from its own equations. The contrastive pre-training loss L1 (Eq. 15) aligns normalized softmax outputs of local and global views; this is a standard self-distillation objective, and the reported OA/AA/Kappa values are not algebraic consequences of that loss. The spiral scan strategy, BPSS block, and Cross Mamba module are architectural components whose contributions are evaluated through ablations (Table II) and parameter studies (Table VII), i.e., empirically. The authors' prior works [4] and [24] are invoked as a benchmark baseline (PiCL) and as guidance for sampling rates, but neither is used to justify the correctness of ECP-Mamba's predictions. The EMA momentum inconsistency and the lack of a spatial buffer in the train/test split are experimental-protocol concerns, not circular reasoning: no equation or fitted constant is renamed as a prediction. The paper is self-contained against external benchmarks, so no circularity score is warranted.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central claim is empirical, so the ledger lists the hyperparameters and assumptions that configure the method rather than derived constants.

free parameters (7)
  • Student temperature tau_s = 0.1
    Sharpens output distribution of local branch in the contrastive loss (Eq. 4).
  • Teacher temperature tau_t = 0.04
    Sharpens target distribution of global branch; chosen by hand.
  • Centering momentum m = 0.9
    EMA update of centering vector C in Eq. 5.
  • EMA momentum lambda = 0.996 (Section VII-A), but Section IV-D says cosine schedule from 0.9995 to 1
    Momentum for teacher parameter update in Eq. 16; text is internally inconsistent.
  • Hidden dimension D = 192
    Model width used in default configuration; selected by parameter analysis in Section VII-D1.
  • BPSS layer counts [Ms, Mt] and M_hat = [1, 1] and 1
    Layer counts selected by ablation and parameter analysis.
  • Patch sizes [k, K] = [16, 32]
    Local and global window sizes; selected by parameter analysis in Section VII-D1.
assumptions (5)
  • standard math State space model discretization via zero-order hold (Eq. 7) is valid for the Mamba update.
    The paper uses standard S4/Mamba discretization formulas.
  • domain assumption The coherence matrix T from Pauli decomposition captures polarimetric scattering information needed for classification.
    The 9-channel input constructed from T's upper triangle (Eq. 21) is assumed sufficient.
  • domain assumption Self-distillation with EMA teacher is a stable training objective that does not collapse.
    The method relies on DINO-style training; collapse is a known risk but not addressed.
  • ad hoc to paper Pixels closer to the central pixel are more causally relevant to the central label.
    This motivates the spiral scan and the tail class token placement (Section IV-C1).
  • domain assumption Randomly sampling a tiny label subset is representative of class distributions.
    With 0.2% labels, some classes have only 2-3 samples; the paper assumes this is sufficient.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ECP-Mamba: An Efficient Multi-scale Self-supervised Contrastive Learning Method with State Space Model for PolSAR Image Classification." pith.science (2026). https://pith.science/paper/6H54NVEM

@misc{pith2026250601040,
  author       = {Pith},
  title        = {Pith review of: ECP-Mamba: An Efficient Multi-scale Self-supervised Contrastive Learning Method with State Space Model for PolSAR Image Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6H54NVEM}},
  note         = {Machine review of arXiv:2506.01040}
}
read the original abstract

Recently, polarimetric synthetic aperture radar (PolSAR) image classification has been greatly promoted by deep neural networks. However,current deep learning-based PolSAR classification methods encounter difficulties due to its dependence on extensive labeled data and the computational inefficiency of architectures like Transformers. This paper presents ECP-Mamba, an efficient framework integrating multi-scale self-supervised contrastive learning with a state space model (SSM) backbone. Specifically, ECP-Mamba addresses annotation scarcity through a multi-scale predictive pretext task based on local-to-global feature correspondences, which uses a simplified self-distillation paradigm without negative sample pairs. To enhance computational efficiency,the Mamba architecture (a selective SSM) is first tailored for pixel-wise PolSAR classification task by designing a spiral scan strategy. This strategy prioritizes causally relevant features near the central pixel, leveraging the localized nature of pixel-wise classification tasks. Additionally, the lightweight Cross Mamba module is proposed to facilitates complementary multi-scale feature interaction with minimal overhead. Extensive experiments across four benchmark datasets demonstrate ECP-Mamba's effectiveness in balancing high accuracy with resource efficiency. On the Flevoland 1989 dataset, ECP-Mamba achieves state-of-the-art performance with an overall accuracy of 99.70%, average accuracy of 99.64% and Kappa coefficient of 99.62e-2. Our code will be available at https://github.com/HaixiaBi1982/ECP_Mamba.

Figures

Figures reproduced from arXiv: 2506.01040 by the authors.

Figure 1
Figure 1. Illustration of ECP-Mamba, which includes Contrastive Pre-training and Classification Fine-tune. Contrastive Pre-training is an unsupervised self [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the scan strategy in traditional Mamba [34] and the proposed spiral scan strategy. Traditional Mamba uses a sequential scan with raster [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. (a) Traditional Mamba without feature fusion. (b) The proposed Cross [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: This figure visually represents the Pauli RGB images and their corresponding ground truth label maps for four PolSAR datasets. (a1) the Pauli RGB [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: (a) Visualization of the proposed spiral scan strategy within a [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Classification results on Flevoland 1989 dataset with different methods. (a) RLRMF-RF. (b) RV-CNN. (c) CV-CNN. (d) PolSF. (e) 3D-GAN. (f) [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Classification results on Flevoland 1991 dataset with different methods. (a) RLRMF-RF. (b) RV-CNN. (c) CV-CNN. (d) PolSF. (e) 3D-GAN. (f) [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Classification results on Oberpfaffenhofen dataset with different methods. (a) RLRMF-RF. (b) RV-CNN. (c) CV-CNN. (d) PolSF. (e) 3D-GAN. (f) [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Classification results on San Francisco dataset with different methods. (a) RLRMF-RF. (b) RV-CNN. (c) CV-CNN. (d) PolSF. (e) 3D-GAN. (f) DB-GC. [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Analysis of OA(%), AA(%) and Kappa(e-2) at various sampling [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

63 extracted references · 44 canonical work pages

  1. [1]

    Classification of sar and polsar images using deep learning: A review,

    H. Parikh, S. Patel, and V . Patel, “Classification of sar and polsar images using deep learning: A review,” International Journal of Image and Data Fusion, vol. 11, no. 1, pp. 1–32, 2020

    work page 2020

  2. [2]

    A polarimetric projection-based scattering characteris- tics extraction tool and its application to polsar image classification,

    W. Han, H. Fu, J. Zhu, S. Zhang, Q. Xie, and J. Hu, “A polarimetric projection-based scattering characteris- tics extraction tool and its application to polsar image classification,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 202, pp. 314–333, 2023

    work page 2023

  3. [3]

    Unsupervised polsar im- age classification using discriminative clustering,

    H. Bi, J. Sun, and Z. Xu, “Unsupervised polsar im- age classification using discriminative clustering,” IEEE Trans. Geosci. Remote Sens. , vol. 55, no. 6, pp. 3531– 3544, 2017

    work page 2017

  4. [4]

    Polarimetry- inspired contrastive learning for class-imbalanced polsar image classification,

    Z. Kuang, H. Bi, F. Li, C. Xu, and J. Sun, “Polarimetry- inspired contrastive learning for class-imbalanced polsar image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 62, p. 5212819, 2024

    work page 2024

  5. [5]

    Meta-graph representation learning for polsar image classification,

    S. Yang, R. Li, Z. Li, H. Meng, Z. Feng, and G. He, “Meta-graph representation learning for polsar image classification,” IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–12, 2024

    work page 2024

  6. [6]

    Polari- metric sar image semantic segmentation with 3d dis- crete wavelet transform and markov random field,

    H. Bi, L. Xu, X. Cao, Y . Xue, and Z. Xu, “Polari- metric sar image semantic segmentation with 3d dis- crete wavelet transform and markov random field,” IEEE Trans. Image Process., vol. 29, pp. 6601–6614, 2020. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, JUNE 2025 17

    work page 2020

  7. [7]

    A review of target de- composition theorems in radar polarimetry,

    S. R. Cloude and E. Pottier, “A review of target de- composition theorems in radar polarimetry,” IEEE Trans. Geosci. Remote Sens., vol. 34, no. 2, pp. 498–518, 1996

    work page 1996

  8. [8]

    A three-component scattering model for polarimetric sar data,

    A. Freeman and S. L. Durden, “A three-component scattering model for polarimetric sar data,” IEEE Trans. Geosci. Remote Sens., vol. 36, no. 3, pp. 963–973, 1998

    work page 1998

Show all 63 references
  1. [9]

    Four-component scattering model for polarimetric sar image decomposition,

    Y . Yamaguchi, T. Moriyama, M. Ishido, and H. Yamada, “Four-component scattering model for polarimetric sar image decomposition,” IEEE Trans. Geosci. Remote Sens., vol. 43, no. 8, pp. 1699–1706, 2005

    2005

  2. [10]

    K- distribution for multi-look processed polarimetric sar imagery,

    J. Lee, D. Schuler, R. Lang, and K. Ranson, “K- distribution for multi-look processed polarimetric sar imagery,” in 1994 IEEE International Geoscience and Remote Sensing Symposium (IGRASS 1994) , vol. 4. IEEE, 1994, pp. 2179–2181

    1994

  3. [11]

    Unsupervised classifi- cation using polarimetric decomposition and the complex wishart classifier,

    J.-S. Lee, M. R. Grunes, T. L. Ainsworth, L.-J. Du, D. L. Schuler, and S. R. Cloude, “Unsupervised classifi- cation using polarimetric decomposition and the complex wishart classifier,” IEEE Trans. Geosci. Remote Sens. , vol. 37, no. 5, pp. 2249–2258, 1999

    1999

  4. [12]

    Classifi- cation with a non-gaussian model for polsar data,

    A. P. Doulgeris, S. N. Anfinsen, and T. Eltoft, “Classifi- cation with a non-gaussian model for polsar data,” IEEE Trans. Geosci. Remote Sens. , vol. 46, no. 10, pp. 2999– 3009, 2008

    2008

  5. [13]

    Support vector machine for multifre- quency sar polarimetric data classification,

    C. Lardeux et al., “Support vector machine for multifre- quency sar polarimetric data classification,” IEEE Trans. Geosci. Remote Sens. , vol. 47, no. 12, pp. 4143–4152, 2009

    2009

  6. [14]

    Tensorial independent component analysis-based feature extraction for polarimetric sar data classification,

    M. Tao, F. Zhou, Y . Liu, and Z. Zhang, “Tensorial independent component analysis-based feature extraction for polarimetric sar data classification,” IEEE Trans. Geosci. Remote Sens. , vol. 53, no. 5, pp. 2481–2495, 2015

    2015

  7. [15]

    Variational bayesian wishart mixture model for polarimetric syn- thetic aperture radar data,

    S. Ren, F. Zhou, W. Fan, and C. Wang, “Variational bayesian wishart mixture model for polarimetric syn- thetic aperture radar data,” in 2019 6th Asia-Pacific Conference on Synthetic Aperture Radar (APSAR), 2019, pp. 1–6

    2019

  8. [16]

    Polarimetric sar image classification using deep convolutional neural networks,

    Y . Zhou, H. Wang, F. Xu, and Y .-Q. Jin, “Polarimetric sar image classification using deep convolutional neural networks,” IEEE Geosci. Remote Sens. Lett. , vol. 13, no. 12, pp. 1935–1939, 2016

    1935

  9. [17]

    Complex- valued convolutional neural network and its application in polarimetric sar image classification,

    Z. Zhang, H. Wang, F. Xu, and Y .-Q. Jin, “Complex- valued convolutional neural network and its application in polarimetric sar image classification,” IEEE Trans. Geosci. Remote Sens. , vol. 55, no. 12, pp. 7177–7188, 2017

    2017

  10. [18]

    Exploring convolutional lstm for polsar image classification,

    L. Wang, X. Xu, H. Dong, R. Gui, R. Yang, and F. Pu, “Exploring convolutional lstm for polsar image classification,” in IGARSS 2018-2018 IEEE International Geoscience and Remote Sensing Symposium . IEEE, 2018, pp. 8452–8455

    2018

  11. [19]

    Polarimetric convolutional network for polsar image classification,

    X. Liu, L. Jiao, X. Tang, Q. Sun, and D. Zhang, “Polarimetric convolutional network for polsar image classification,” IEEE Transactions on Geoscience and Remote Sensing, vol. 57, no. 5, pp. 3040–3054, 2019

    2019

  12. [20]

    Exploring vision transformers for polarimetric sar image classification,

    H. Dong, L. Zhang, and B. Zou, “Exploring vision transformers for polarimetric sar image classification,” IEEE Trans. Geosci. Remote Sens. , vol. 60, pp. 1–15, 2021

    2021

  13. [21]

    Local window attention transformer for polarimetric sar image classification,

    A. Jamali, S. K. Roy, A. Bhattacharya, and P. Ghamisi, “Local window attention transformer for polarimetric sar image classification,” IEEE Geoscience and Remote Sensing Letters, vol. 20, pp. 1–5, 2023

    2023

  14. [22]

    Polarimetric sar image classification based on hierarchical scattering-spatial in- teraction transformer,

    J. Geng, Y . Zhang, and W. Jiang, “Polarimetric sar image classification based on hierarchical scattering-spatial in- teraction transformer,” IEEE Transactions on Geoscience and Remote Sensing , 2024

    2024

  15. [23]

    Semi-supervised complex-valued gan for polarimetric sar image classification,

    Q. Sun, X. Li, L. Li, X. Liu, F. Liu, and L. Jiao, “Semi-supervised complex-valued gan for polarimetric sar image classification,” in 2019 IEEE International Geoscience and Remote Sensing Symposium (IGRASS 2019), 2019, pp. 3245–3248

    2019

  16. [24]

    Dual-branch polsar image classification based on graphmae and local feature extraction,

    Y . Wang, Z. Guo, H. Bi, D. Hong, and C. Xu, “Dual-branch polsar image classification based on graphmae and local feature extraction,” arXiv preprint arXiv:2408.04294, 2024

    2024 arXiv

  17. [25]

    Polsar image clas- sification with complex-valued diffusion model as repre- sentation learners,

    Z. Kuang, K. Liu, H. Bi, and F. Li, “Polsar image clas- sification with complex-valued diffusion model as repre- sentation learners,” IEEE Transactions on Aerospace and Electronic Systems, pp. 1–21, 2025

    2025

  18. [26]

    Unsupervised deep representation learning and few-shot classification of polsar images,

    L. Zhang, S. Zhang, B. Zou, and H. Dong, “Unsupervised deep representation learning and few-shot classification of polsar images,” IEEE Trans. Geosci. Remote Sens. , vol. 60, pp. 1–16, 2022

    2022

  19. [27]

    A mutual information-based self-supervised learning model for polsar land cover classification,

    B. Ren, Y . Zhao, B. Hou, J. Chanussot, and L. Jiao, “A mutual information-based self-supervised learning model for polsar land cover classification,” IEEE Trans. Geosci. Remote Sens., vol. 59, no. 11, pp. 9224–9237, 2021

    2021

  20. [28]

    Exploring polsar images representation via self-supervised learning and its appli- cation on few-shot classification,

    W. Zhang, Z. Pan, and Y . Hu, “Exploring polsar images representation via self-supervised learning and its appli- cation on few-shot classification,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022

    2022

  21. [29]

    A graph-based semisupervised deep learning model for polsar image classification,

    H. Bi, J. Sun, and Z. Xu, “A graph-based semisupervised deep learning model for polsar image classification,” IEEE Trans. Geosci. Remote Sens. , vol. 57, no. 4, pp. 2116–2132, 2019

    2019

  22. [30]

    Recent advances in deep learning models: a systematic literature review,

    R. Malhotra and P. Singh, “Recent advances in deep learning models: a systematic literature review,” pp. 44 977–45 060, 2023

    2023

  23. [31]

    Modeling polsar classification using convolutional neural network with homogeneity based kernel selection,

    H. Parikh, S. Patel, and V . Patel, “Modeling polsar classification using convolutional neural network with homogeneity based kernel selection,” Modeling Earth Systems and Environment, vol. 9, no. 4, pp. 3801–3813, 2023

    2023

  24. [32]

    Vision mamba: Efficient visual representation learning with bidirectional state space model,

    L. Zhu, B. Liao, Q. Zhang, X. Wang, W. Liu, and X. Wang, “Vision mamba: Efficient visual representation learning with bidirectional state space model,” arXiv preprint arXiv:2401.09417, 2024

    2024 arXiv

  25. [33]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural informa- tion processing systems , vol. 30, 2017

    2017

  26. [34]

    Mamba: Linear-time sequence modeling with selective state spaces,

    A. Gu and T. Dao, “Mamba: Linear-time sequence modeling with selective state spaces,” arXiv preprint arXiv:2312.00752, 2023. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, JUNE 2025 18

    2023 arXiv

  27. [35]

    Efficiently modeling long sequences with structured state spaces,

    A. Gu, K. Goel, and C. R ´e, “Efficiently modeling long sequences with structured state spaces,” arXiv preprint arXiv:2111.00396, 2021

    2021 arXiv

  28. [36]

    Visual mamba: A survey and new outlooks,

    R. Xu, S. Yang, Y . Wang, Y . Cai, B. Du, and H. Chen, “Visual mamba: A survey and new outlooks,” 2024. [Online]. Available: https://arxiv.org/abs/2404.18861

    2024 arXiv

  29. [37]

    Intensity and phase statistics of multilook polarimetric and inter- ferometric sar imagery,

    J.-S. Lee, K. Hoppel, S. Mango, and A. Miller, “Intensity and phase statistics of multilook polarimetric and inter- ferometric sar imagery,” IEEE Trans. Geosci. Remote Sens., vol. 32, no. 5, pp. 1017–1028, 1994

    1994

  30. [38]

    A review of target decomposi- tion theorems in radar polarimetry,

    S. Cloude and E. Pottier, “A review of target decomposi- tion theorems in radar polarimetry,” IEEE Trans. Geosci. Remote Sens., vol. 34, no. 2, pp. 498–518, 1996

    1996

  31. [39]

    Random neighbor pixel-block-based deep re- current learning for polarimetric sar image classification,

    J. Ni, F. Zhang, Q. Yin, Y . Zhou, H.-C. Li, and W. Hong, “Random neighbor pixel-block-based deep re- current learning for polarimetric sar image classification,” IEEE Transactions on Geoscience and Remote Sensing , vol. 59, no. 9, pp. 7557–7569, 2020

    2020

  32. [40]

    An active deep learning approach for minimally supervised polsar image classification,

    H. Bi, F. Xu, Z. Wei, Y . Xue, and Z. Xu, “An active deep learning approach for minimally supervised polsar image classification,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 11, pp. 9378–9395, 2019

    2019

  33. [41]

    Unsupervised feature learning via non-parametric instance discrimina- tion,

    Z. Wu, Y . Xiong, S. X. Yu, and D. Lin, “Unsupervised feature learning via non-parametric instance discrimina- tion,” in 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2018) , 2018, pp. 3733– 3742

    2018

  34. [42]

    Mo- mentum contrast for unsupervised visual representation learning,

    K. He, H. Fan, Y . Wu, S. Xie, and R. Girshick, “Mo- mentum contrast for unsupervised visual representation learning,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2020) , 2020, pp. 9726–9735

    2020

  35. [43]

    What makes for good views for contrastive learning?

    Y . Tian, C. Sun, B. Poole, D. Krishnan, C. Schmid, and P. Isola, “What makes for good views for contrastive learning?” Advances in neural information processing systems, vol. 33, pp. 6827–6839, 2020

    2020

  36. [44]

    Bootstrap your own latent-a new approach to self-supervised learning,

    J.-B. Grill et al. , “Bootstrap your own latent-a new approach to self-supervised learning,” in 2020 Advances in Neural Information Processing Systems (NIPS 2020) , vol. 33, 2020, pp. 21 271–21 284

    2020

  37. [45]

    Emerging properties in self-supervised vision transformers,

    M. Caron, H. Touvron, I. Misra, H. J ´egou, J. Mairal, P. Bojanowski, and A. Joulin, “Emerging properties in self-supervised vision transformers,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 9650–9660

    2021

  38. [46]

    Multi-scale con- trastive learning method for polsar image classification,

    W. Hua, C. Wang, N. Sun, and L. Liu, “Multi-scale con- trastive learning method for polsar image classification,” Journal of Applied Remote Sensing , vol. 18, no. 1, pp. 014 502–014 502, 2024

    2024

  39. [47]

    Vmamba: Visual state space model,

    Y . Liu, Y . Tian, Y . Zhao, H. Yu, L. Xie, Y . Wang, Q. Ye, and Y . Liu, “Vmamba: Visual state space model,” 2024. [Online]. Available: https://arxiv.org/abs/2401.10166

    2024 arXiv

  40. [48]

    Vision mamba: A comprehensive survey and taxonomy,

    X. Liu, C. Zhang, and L. Zhang, “Vision mamba: A comprehensive survey and taxonomy,” arXiv preprint arXiv:2405.04404, 2024

    2024 arXiv

  41. [49]

    Plainmamba: Improving non- hierarchical mamba in visual recognition,

    C. Yang, Z. Chen, M. Espinosa, L. Ericsson, Z. Wang, J. Liu, and E. J. Crowley, “Plainmamba: Improving non- hierarchical mamba in visual recognition,” arXiv preprint arXiv:2403.17695, 2024

    2024 arXiv

  42. [50]

    Localmamba: Visual state space model with windowed selective scan,

    T. Huang, X. Pei, S. You, F. Wang, C. Qian, and C. Xu, “Localmamba: Visual state space model with windowed selective scan,” arXiv preprint arXiv:2403.09338 , 2024

    2024 arXiv

  43. [51]

    Mambaad: Ex- ploring state space models for multi-class unsupervised anomaly detection,

    H. He, Y . Bai, J. Zhang, Q. He, H. Chen, Z. Gan, C. Wang, X. Li, G. Tian, and L. Xie, “Mambaad: Ex- ploring state space models for multi-class unsupervised anomaly detection,” arXiv preprint arXiv:2404.06564 , 2024

    2024 arXiv

  44. [52]

    Efficientvmamba: Atrous selective scan for light weight visual mamba,

    X. Pei, T. Huang, and C. Xu, “Efficientvmamba: Atrous selective scan for light weight visual mamba,” arXiv preprint arXiv:2403.09977, 2024

    2024 arXiv

  45. [53]

    Rsmamba: Remote sensing image classification with state space model,

    K. Chen, B. Chen, C. Liu, W. Li, Z. Zou, and Z. Shi, “Rsmamba: Remote sensing image classification with state space model,” IEEE Geoscience and Remote Sens- ing Letters, 2024

    2024

  46. [54]

    Spectral- mamba: Efficient mamba for hyperspectral image clas- sification,

    J. Yao, D. Hong, C. Li, and J. Chanussot, “Spectral- mamba: Efficient mamba for hyperspectral image clas- sification,” arXiv preprint arXiv:2404.08489 , 2024

    2024 arXiv

  47. [55]

    Ssumamba: Spatial- spectral selective state space model for hyperspectral image denoising,

    G. Fu, F. Xiong, J. Lu, and J. Zhou, “Ssumamba: Spatial- spectral selective state space model for hyperspectral image denoising,” IEEE Transactions on Geoscience and Remote Sensing, 2024

    2024

  48. [56]

    Changemamba: Remote sensing change detection with spatio-temporal state space model,

    H. Chen, J. Song, C. Han, J. Xia, and N. Yokoya, “Changemamba: Remote sensing change detection with spatio-temporal state space model,” arXiv preprint arXiv:2404.03425, 2024

    2024 arXiv

  49. [57]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems, vol. 33, pp. 6840–6851, 2020

    2020

  50. [58]

    Decoupled weight decay regularization,

    I. Loshchilov, “Decoupled weight decay regularization,” arXiv preprint arXiv:1711.05101 , 2017

    2017 arXiv

  51. [59]

    Polyloss: A polynomial expansion perspective of classification loss functions,

    Z. Leng, M. Tan, C. Liu, E. D. Cubuk, X. Shi, S. Cheng, and D. Anguelov, “Polyloss: A polynomial expansion perspective of classification loss functions,” arXiv preprint arXiv:2204.12511 , 2022

    2022 arXiv

  52. [60]

    SGDR: stochastic gradient descent with restarts,

    I. Loshchilov and F. Hutter, “SGDR: stochastic gradient descent with restarts,” CoRR, vol. abs/1608.03983, 2016. [Online]. Available: http://arxiv.org/abs/1608.03983

    2016 arXiv

  53. [61]

    Polsar image classification based on robust low-rank feature ex- traction and markov random field,

    H. Bi, J. Yao, Z. Wei, D. Hong, and J. Chanussot, “Polsar image classification based on robust low-rank feature ex- traction and markov random field,” IEEE Geosci. Remote Sens. Lett., vol. 19, pp. 1–5, 2022

    2022

  54. [62]

    Polsar image classification based on 3d generative adversarial networks,

    A. Jamali, F. Mohammadimanesh, M. Mahdianpari, B. Lu, and A. Bhattacharya, “Polsar image classification based on 3d generative adversarial networks,” in 2023 IEEE India Geoscience and Remote Sensing Symposium (InGARSS). IEEE, 2023, pp. 1–4

    2023

  55. [63]

    Polsf: Polsar image datasets on san francisco,

    X. Liu, L. Jiao, F. Liu, D. Zhang, and X. Tang, “Polsf: Polsar image datasets on san francisco,” in Intelligence Science IV, Z. Shi, Y . Jin, and X. Zhang, Eds. Cham: Springer International Publishing, 2022, pp. 214–219

    2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.