Pith. sign in

REVIEW 4 major objections 6 minor 40 references

Direction-adaptive Mamba: Spatial-Frequency Dual-Domain Collaborative Learning for PolSAR Image Classification

T0 review · 4 major / 6 minor · reviewed 2026-07-30 · grok-4.5

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

Pith's one-line read Direction-adaptive Mamba plus spatial-frequency fusion beats prior PolSAR classifiers by following edge orientations instead of fixed scans.

desk verdict Solid engineering combo of oriented Mamba on NSCT bands plus spatial Mamba, but the SOTA and attribution claims rest on sub-0.4% single-split margins that the authors’ own ablations undercut. read the letter →

arxiv 2607.23464 v1 pith:WAPQ3KEQ submitted 2026-07-26 eess.IV cs.CV

Junfei Shi , Yu Cheng , Haojia Zhang , Wenqiang Hua , Junhuai Li , Maoguo Gong This is my paper · ORCID
classification eess.IVcs.CV
keywords PolSARclassificationMambadirection-adaptivescanningNSCTspatial-frequencyfusionanisotropicscatteringstatespacemodel
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

Polarimetric SAR images encode how surfaces scatter radar energy in different orientations, so land-cover classes often differ by thin, direction-sensitive edges rather than by smooth blobs. Existing Mamba models either stay purely spatial and blur those edges, or use fixed raster scans that ignore the preferred direction of each texture. This paper claims that decomposing the image with a non-subsampled contourlet transform, then letting a shared Mamba scan each high-frequency subband along its own edge orientation, while a second spatial Mamba keeps raw scattering context, and a gate fuses the three streams, yields more accurate class maps. On standard benchmarks the full pipeline reports higher overall accuracy and cleaner boundaries than recent CNN, Transformer, diffusion, and Mamba baselines. A sympathetic reader cares because the same linear-complexity backbone can now respect the anisotropic physics of PolSAR without quadratic attention cost.

What carries the argument

Direction-adaptive Mamba (DA-Mamba): each NSCT high-frequency subband is serialized by projecting pixels onto that subband’s principal angle, then processed by a shared selective state-space model so long-range context flows along the true edge orientation rather than a fixed raster.

What would settle it

Re-run the exact same training protocol with multiple independent 5 percent draws (or published multi-run means and standard deviations) on Xi’an and Oberpfaffenhofen; if the mean OA of DA-Mamba no longer exceeds NGDiffSM and the other listed baselines, the central superiority claim fails.

Watch

Extended reading notes

Core claim

The authors establish that an edge-aligned, direction-adaptive Mamba operating on NSCT high-frequency subbands, paired with a spatial 2D Mamba on raw polarimetry and a gated dual-domain fusion, produces more discriminative PolSAR features and higher classification accuracy than fixed-scan or single-domain Mamba and than listed state-of-the-art methods.

Load-bearing premise

A single random 5 percent labeled split is treated as a stable enough basis to claim superiority even when absolute gains over the strongest baseline are a few tenths of a percent and the full model is not best on every dataset in the authors’ own ablations.

Editorial extensions

If this is right

  • PolSAR classifiers can keep Mamba’s linear complexity while still modeling anisotropic scattering and weak boundaries.
  • NSCT multi-directional high-frequency subbands become a natural substrate for orientation-aware state-space scanning rather than only for CNN or wavelet front-ends.
  • Gated fusion of raw spatial scattering with low- and high-frequency contourlet streams is a reusable pattern for other dual-domain remote-sensing tasks.
  • Edge-aligned serialization can replace fixed four-direction 2D scans whenever the data have known preferred orientations.

Reading between the lines

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

  • The same orientation-aware serialization could be dropped into other multi-scale decompositions (shearlets, dual-tree wavelets) without redesigning the Mamba core.
  • Because margins are small and few-shot generalization is already flagged as weak, the practical next test is whether the direction-adaptive prior still helps under 1 percent labels or cross-sensor transfer.
  • If the gain is truly from matching scan direction to scattering anisotropy, similar adaptive scans should help single-pol SAR and optical edge-heavy segmentation where fixed Mamba scans currently dominate.
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 / 6 minor

Summary. The manuscript proposes DA-Mamba-DDCL, a PolSAR image classification framework that couples (i) a spatial-domain 2D Mamba over raw polarimetric features with (ii) an NSCT-based frequency branch in which high-frequency directional subbands are serialized along their decomposition orientation (ζ_k(i,j) = i·cosθ_k + j·sinθ_k) and processed by a shared Mamba SSM, plus a lightweight low-frequency branch and a gated three-stream fusion. The authors report state-of-the-art results on four PolSAR benchmarks (Xi'an OA 97.17%, Oberpfaffenhofen OA 95.52%, with San Francisco and Flevoland relegated to supplementary), module ablations (Table 3), and sensitivity analyses over NSCT direction count, Mamba state dimension, and training ratio.

Significance. The direction-aware serialization idea is a clean and well-motivated match to NSCT structure: aligning the SSM scan order with each subband's principal orientation is a parameter-cheap modification with a clear physical rationale for anisotropic PolSAR scattering, and the SSF multi-directional fusion is a sensible complement. The paper is stronger than many architecture papers in this space in shipping a four-dataset evaluation, a five-configuration module ablation, and three hyperparameter sensitivity studies, and the authors commit to releasing code, which would make the pipeline (rotated-Sobel edge enhancement, permutation-based scans, gated fusion) reproducible in principle. If the reported gains are statistically robust, this would be a useful contribution to the growing Mamba-for-remote-sensing literature. However, as detailed below, the decisive margins over the strongest baseline are smaller than the run-to-run and hyperparameter variance the paper itself documents, so the central empirical claim is currently under-supported.

major comments (4)
  1. [Experimental Setup / Tables 1–2] The headline superiority claim rests on a single random 5%/1% train/val split with no repeated runs, no standard deviations, and no significance testing. The decisive margins over the strongest baseline (NGDiffSM) are +0.18% OA on Xi'an and +0.38% OA on Oberpfaffenhofen. The paper's own sensitivity studies show variance of this order or larger from a single hyperparameter change: Table 4 shows direction count swinging Xi'an OA from 95.26% (4 directions) to 97.17% (8) to 95.48% (16), and Fig. 6 shows d_state moving Xi'an OA between 96.02% and 97.17%. Patch-based PolSAR classifiers at 5% labels routinely exhibit seed-to-seed OA fluctuations of several tenths of a percent. As reported, the SOTA margins are indistinguishable from seed noise. Please report mean ± std over at least 5 random splits/seeds for the proposed method and the top baselines, or an equivalent significance test.
  2. [Ablation Analysis, Table 3] The ablation results do not consistently support contribution 2 (the dual-domain collaborative design). The full five-module model is not the best configuration on San Francisco (OA 99.55 vs. 99.70 for the two-branch variant) or on Oberpfaffenhofen (OA 95.52 vs. 95.83 for the variant without one module). The text says only that the full model 'remains competitive' there, but the abstract and contribution 2 claim that dual-domain fusion 'strengthens discriminability' — a claim that requires the complete design to be at least non-inferior across datasets. Combined with Major Comment 1 (single split), these reversals may be noise, but then so may the gains. Either the ablation needs multi-run statistics, or the claims need to be qualified to 'best on Xi'an and Flevoland' with an honest discussion of where modules hurt.
  3. [Experimental Setup / Results] The protocol for baseline numbers is not stated in the main text: it is unclear whether DFGCN, HybridCVNet, CV-MsAtViT, SpectralNet, S2Mamba, and NGDiffSM were re-trained under the identical 5%/1% split, patch size, and augmentation, or whether numbers were transcribed from prior publications under different protocols. This matters because SpectralNet and S2Mamba are hyperspectral methods whose adaptation to PolSAR inputs requires non-trivial choices, and because S2Mamba's Oberpfaffenhofen results (OA 74.91%, road class 8.06%) suggest a possible training failure rather than a fair comparison. Please state explicitly in the main text (not only the supplementary) that all baselines were re-run under the identical protocol, and describe the input adaptation for the HSI-derived baselines.
  4. [Experimental Setup] With 64×64 patches and random pixel-level train/test partitioning, spatially overlapping patches straddling the train/test boundary are likely, which is known to inflate OA in patch-based PolSAR/HSI classification and can easily account for sub-0.5% effects of the size claimed here. The manuscript does not state whether any spatial-disjointness measure (buffer zones, block-wise splits, or non-overlapping patch extraction) was used. Please clarify the sampling procedure and, if overlapping patches are used, either add a spatially disjoint evaluation or explicitly acknowledge this as a limitation shared with the compared protocols.
minor comments (6)
  1. [Abstract] The abstract states 'Evaluated on three real-world PolSAR datasets,' but the Experimental Setup evaluates four (Xi'an, San Francisco, Flevoland, Oberpfaffenhofen). Please reconcile.
  2. [References] Several references are duplicated with 'a'/'b' suffixes pointing to identical entries: Dong, Zhang, and Zou 2021a/2021b; Freeman and Durden 1998a/1998b; Zhou et al. 2016a/2016b; Lee, Grunes, and Kwok 1994a/1994b. Also, Yu et al. 2015 ('Learning cross space mapping via DNN using large scale click-through logs') is a web-search paper and is almost certainly the wrong citation for CNN-based PolSAR classification.
  3. [Proposed Method, Multi-directional subbands fusion] The SSF fusion equation introduces Softmax in the text ('Softmax(·) performs weight normalization') but the displayed formula F_hf^fused = Σ_k C(F_hf^(k)) ⊙ F_hf^(k) contains no normalization; please make the equation match the description. Similarly, the terminology switches between 'HF' and 'HL' subbands.
  4. [Conclusion] The conclusion claims superiority in 'computational efficiency,' but no FLOPs, parameter counts, or runtime comparisons appear anywhere in the manuscript. Either add an efficiency table or remove the claim.
  5. [Proposed Method / Experimental Setup] The polarimetric input X_ori is never fully specified: the number of channels C_ori and whether the input is the scattering vector, coherency/covariance matrix elements, or a decomposition feature set should be stated in the main text rather than deferred to the supplementary.
  6. [Throughout] Typos and figure issues: 'muti-directional'; Fig. 6 is referenced as 'Fig. 6(b)' though no panels are described; 'data sets' (Table 3 caption) vs 'datasets'; inconsistent capitalization of 'Direction-adaptive Mamba' vs 'DA-Mamba' throughout.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: supervised architecture paper; claimed gains are empirical benchmark results, not definitional or self-citation-forced derivations.

full rationale

DA-Mamba-DDCL is a standard end-to-end supervised PolSAR classifier. The method chain is architectural composition (NSCT decomposition → edge-guided HF enhancement → direction-aware serialization + shared Mamba → LF conv branch → spatial 2D Mamba → gated dual-domain fusion → cross-entropy), not a first-principles derivation that could collapse into its inputs. Training uses ordinary supervised CE loss on labeled patches; reported OA/AA/Kappa are held-out classification metrics against third-party and one overlapping-author baseline (NGDiffSM), not quantities fitted then re-labeled as predictions. NSCT, Sobel, Mamba, and gating are adopted as design choices with ablations, not uniqueness theorems or self-cited ansatze that forbid alternatives. Mild author overlap on the NGDiffSM baseline is ordinary competitive comparison and does not underwrite the architecture by construction. Statistical concerns (single split, sub-percent margins, ablation non-monotonicity) affect evidential strength, not circularity. No step reduces a claimed result to its own definition or fit.

Assumptions & free parameters 6 free parameters · 5 assumptions · 4 invented entities

The claim rests on standard deep-learning practice plus domain choices about PolSAR representation: NSCT as the right multi-directional front-end, projection sorting as a faithful edge-aligned scan, four-way 2D Mamba as adequate spatial context, and a single 5% labeled split as a sufficient test of superiority. Free parameters are the usual training and architecture knobs (K, d_state, α, lr, etc.). Invented entities are architectural modules, not physical objects; none have independent evidence outside this paper’s ablations.

free parameters (6)
  • NSCT direction count K = 8
    Chosen by ablation among {4,8,16}; K=8 fixed as default because it maximized OA on Xi’an/Flevoland.
  • Mamba state dimension d_state = 32
    Selected from {8,16,32,64} via validation-style sweep; 32 chosen for Xi’an peak OA.
  • edge enhancement mix α = learnable (unreported)
    Learnable scalar controlling Sobel-based additive enhancement on HF subbands; value not reported.
  • training label ratio = 5%
    Primary tables use 5% train / 1% val after sweeping 1–10%; protocol choice that strongly affects absolute OA.
  • AdamW lr and weight decay = 1e-4, 1e-3
    Hand-set optimization hyperparameters used for all reported runs.
  • patch size / batch / epochs = 64 / 128 / 300
    Training schedule knobs fixed without broader sensitivity beyond train-ratio plot.
assumptions (5)
  • domain assumption NSCT shift-invariant multi-directional subbands are an appropriate and sufficient frequency front-end for anisotropic PolSAR structure.
    Invoked in ‘DA-Mamba for Frequency Domain’ as the decomposition that separates global LF context from oriented HF edges; not compared head-to-head with wavelets/shearlets in the main text.
  • ad hoc to paper Serializing each HF subband by projection ζ_k(i,j)=i cos θ_k + j sin θ_k aligns SSM state propagation with true scattering anisotropy better than fixed 2D scans.
    Core design hypothesis of Direction-adaptive Mamba (Fig. 3 and serialization equations); supported only by within-paper ablations replacing Dir-Mamba with conventional 2D Mamba.
  • domain assumption A single random draw of 5% labeled pixels is a fair protocol for ranking methods against published SOTA.
    Stated under Experimental Setup as ‘following common PolSAR classification protocols’; load-bearing for the superiority claim given tiny margins.
  • domain assumption Selective state-space (Mamba) sequence modeling with linear complexity is a valid global context backbone for PolSAR patches.
    Taken from Gu & Dao 2023 and prior remote-sensing Mamba works cited in Related Work; assumed rather than re-derived.
  • ad hoc to paper Gated channel-concat fusion of spatial, LF, and HF streams yields more discriminative features than fixed fusion.
    Dual-domain Feature Fusion section; justified by end-task gains, not by an external theory of optimal fusion.
invented entities (4)
  • Direction-adaptive Mamba (DA-Mamba) with edge-aligned serialization
    purpose: Scan each NSCT high-frequency subband along its principal orientation before shared SSM modeling to capture anisotropic edges.
    Central named module; defined via π_k sorting and inverse restore. Evidence is only internal accuracy/ablation tables.
  • Edge-guided HF enhancement (rotated Sobel + learnable α)
    purpose: Inject directional edge cues from raw PolSAR into NSCT HF features before Mamba.
    Introduced under ‘Edge-guided feature Enhancement’; ablation ‘without Ori-Guided Enhancement’ is the only external check.
  • Spatial Selective Fusion (SSF) for multi-directional HF subbands
    purpose: Weight and merge K oriented HF Mamba outputs into one HF stream.
    Defined as sum_k C(F_hf^(k)) ⊙ F_hf^(k); no independent validation outside the full pipeline.
  • DA-Mamba-DDCL gated dual-domain collaborative framework
    purpose: Jointly learn raw spatial scattering, LF global structure, and HF directional texture for classification.
    Overall system name in contributions; performance is the sole warrant.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Direction-adaptive Mamba: Spatial-Frequency Dual-Domain Collaborative Learning for PolSAR Image Classification." pith.science (2026). https://pith.science/paper/WAPQ3KEQ

@misc{pith2026260723464,
  author       = {Pith},
  title        = {Pith review of: Direction-adaptive Mamba: Spatial-Frequency Dual-Domain Collaborative Learning for PolSAR Image Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WAPQ3KEQ}},
  note         = {Machine review of arXiv:2607.23464}
}
read the original abstract

Deep learning dominates polarimetric synthetic aperture radar (PolSAR) image classification, with Mamba architectures serving as favorable backbones due to linear complexity and strong global modeling capacity. However, existing PolSAR Mamba methods have two critical flaws: pure spatial processing discards fine-grained edges and textures, and fixed scanning patterns fail to model direction-variant anisotropic scattering and weak boundaries essential for PolSAR physical analysis. This work proposes DA-Mamba, a direction-adaptive Mamba framework with dual-domain collaborative learning for PolSAR classification. Equipped with an edge-aligned direction-adaptive scanning scheme, DA-Mamba captures long-range spatial dependencies and accurate boundary details. It adopts the Non-Subsampled Contourlet Transform (NSCT) to separate PolSAR data into low-frequency global components and multi-directional high-frequency subbands, extracting anisotropic structural features from high-frequency information while preserving global context via low-frequency branches. A dual-domain collaborative learning module further integrates spatial scattering and frequency-domain representations to strengthen feature discriminability. Evaluated on three real-world PolSAR datasets, DA-Mamba surpasses state-of-the-art methods, verifying the efficacy of the proposed adaptive scanning and dual-domain fusion designs. Code will be publicly available.

Figures

Figures reproduced from arXiv: 2607.23464 by the authors.

Figure 1
Figure 1. Comparison of conventional and proposed meth [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Framework of the proposed method [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the proposed direction-aware serial [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison on the Xi’an dataset. (a) [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison on the Oberpfaffenhofen [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 7
Figure 7. Figure 7: Effect of the training sample ratio on classification [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 6
Figure 6. Figure 6: Effect of the Mamba state dimension on classifica [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 4 linked inside Pith

  1. [1]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Antropov, Oleg and Rauste, Yrjö and Astola, Heikki and Praks, Jaan and Häme, Tuomas and Hallikainen, Martti T , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2013 , type =

    2013

  2. [2]

    arXiv preprint arXiv:2312.00752 , year =

    Gu, Albert and Dao, Tri , title =. arXiv preprint arXiv:2312.00752 , year =

    arXiv

  3. [3]

    arXiv preprint arXiv:2111.00396 , year =

    Gu, Albert and Goel, Karan and Ré, Christopher , title =. arXiv preprint arXiv:2111.00396 , year =

    arXiv

  4. [4]

    Remote Sensing , volume =

    Ji, Yaqi and Sri Sumantyo, Josaphat Tetuko and Chua, Ming Yam and Waqar, Mirza Muhammad , title =. Remote Sensing , volume =. 2018 , type =

    2018

  5. [5]

    arXiv preprint arXiv:2208.04933 , year =

    Smith, Jimmy TH and Warrington, Andrew and Linderman, Scott W , title =. arXiv preprint arXiv:2208.04933 , year =

    arXiv

  6. [6]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =

    Wiseman, Grant and McNairn, Heather and Homayouni, Saeid and Shang, Jiali , title =. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =. 2014 , type =

    2014

  7. [7]

    Remote Sensing , volume =

    Xiang, Deliang and Tang, Tao and Hu, Canbin and Fan, Qinghui and Su, Yi , title =. Remote Sensing , volume =. 2016 , type =

    2016

  8. [8]

    International Journal of Remote Sensing , volume =

    Lee, Jong-Sen and Grunes, Mitchell R and Kwok, Ronald , title =. International Journal of Remote Sensing , volume =. 1994 , type =

    1994

Show all 40 references
  1. [9]

    IEEE transactions on geoscience and remote sensing , volume =

    Freeman, Anthony and Durden, Stephen L , title =. IEEE transactions on geoscience and remote sensing , volume =. 1998 , type =

    1998

  2. [10]

    2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) , publisher =

    Guo, Shenglong and Tian, Yurun and Li, Yang and Chen, Shiqiang and Hong, Wen , title =. 2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) , publisher =. 2015 , type =

    2015

  3. [11]

    IEEE Geoscience and Remote Sensing Letters , volume =

    Zhou, Yu and Wang, Haipeng and Xu, Feng and Jin, Ya-Qiu , title =. IEEE Geoscience and Remote Sensing Letters , volume =. 2016 , type =

    2016

  4. [12]

    IEEE Transactions on Multimedia , volume =

    Yu, Wei and Yang, Kuiyuan and Bai, Yalong and Yao, Hongxun and Rui, Yong , title =. IEEE Transactions on Multimedia , volume =. 2015 , type =

    2015

  5. [13]

    Remote Sensing , volume =

    Wang, Hongmiao and Xing, Cheng and Yin, Junjun and Yang, Jian , title =. Remote Sensing , volume =. 2022 , type =

    2022

  6. [14]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Dong, Hongwei and Zhang, Lamei and Zou, Bin , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2021 , type =

    2021

  7. [15]

    IEEE Transactions on Geoscience and Remote Sensing , ISSN =

    Kuang, Zuzheng and Bi, Haixia and Li, Fan and Xu, Chen , title =. IEEE Transactions on Geoscience and Remote Sensing , ISSN =. 2025 , type =

    2025

  8. [16]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Wang, Guanchun and Zhang, Xiangrong and Peng, Zelin and Zhang, Tianyang and Jiao, Licheng , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2025 , type =

    2025

  9. [17]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    He, Yan and Tu, Bing and Liu, Bo and Li, Jun and Plaza, Antonio , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2024 , type =

    2024

  10. [18]

    International Journal of Remote Sensing , volume =

    Cui, Xiandai and Zhang, Li , title =. International Journal of Remote Sensing , volume =. 2025 , type =

    2025

  11. [19]

    Digital Signal Processing , pages =

    An, Wenlong and Deng, Wanyu and Yao, Yongsheng and Zong, Hui and Jiang, Dongshan , title =. Digital Signal Processing , pages =. 2026 , type =

    2026

  12. [20]

    IEEE Geoscience and Remote Sensing Letters , volume =

    Chen, Keyan and Chen, Bowen and Liu, Chenyang and Li, Wenyuan and Zou, Zhengxia and Shi, Zhenwei , title =. IEEE Geoscience and Remote Sensing Letters , volume =. 2024 , type =

    2024

  13. [21]

    Remote Sensing , volume =

    Ding, Hao and Xia, Bo and Liu, Weilin and Zhang, Zekai and Zhang, Jinglin and Wang, Xing and Xu, Sen , title =. Remote Sensing , volume =. 2024 , type =

    2024

  14. [22]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , ISSN =

    Zheng, Kang and Yu, Minmin and Liu, Zhifei and Bao, Shuai and Pan, Ziwu and Song, Yongze and Zhu, Lianqi and Xie, Zunyi , title =. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , ISSN =. 2025 , type =

    2025

  15. [23]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Lee, Jong-Sen and Grunes, Mitchell R and Ainsworth, Thomas L and Du, Li-Jen and Schuler, Dale L and Cloude, Shane R , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 1999 , type =

    1999

  16. [24]

    IEEE transactions on geoscience and remote sensing , volume =

    Cloude, Shane R and Pottier, Eric , title =. IEEE transactions on geoscience and remote sensing , volume =. 1997 , type =

    1997

  17. [25]

    IEEE Transactions on geoscience and remote sensing , volume =

    Yamaguchi, Yoshio and Moriyama, Toshifumi and Ishido, Motoi and Yamada, Hiroyoshi , title =. IEEE Transactions on geoscience and remote sensing , volume =. 2005 , type =

    2005

  18. [26]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Singh, Gulab and Yamaguchi, Yoshio and Park, Sang-Eun , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2012 , type =

    2012

  19. [27]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Liu, Fang and Jiao, Licheng and Hou, Biao and Yang, Shuyuan , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2016 , type =

    2016

  20. [28]

    IEEE Transactions on Geoscience and Remote Sensing , volume =

    Cui, Yuanhao and Liu, Fang and Jiao, Licheng and Guo, Yuwei and Liang, Xuefeng and Li, Lingling and Yang, Shuyuan and Qian, Xiaoxue , title =. IEEE Transactions on Geoscience and Remote Sensing , volume =. 2021 , type =

    2021

  21. [29]

    Remote Sensing , volume =

    Dong, Hongwei and Zhang, Lamei and Zou, Bin , title =. Remote Sensing , volume =. 2020 , type =

    2020

  22. [30]

    IEEE Geoscience and Remote Sensing Letters , volume =

    Ding, Lei and Zheng, Kai and Lin, Dong and Chen, Yuxing and Liu, Bing and Li, Jiansheng and Bruzzone, Lorenzo , title =. IEEE Geoscience and Remote Sensing Letters , volume =. 2021 , type =

    2021

  23. [31]

    IEEE Geoscience and Remote Sensing Letters , volume =

    Alkhatib, Mohammed Q , title =. IEEE Geoscience and Remote Sensing Letters , volume =. 2024 , type =

    2024

  24. [32]

    Pattern Recognition , volume =

    Li, Lingling and Ma, Liyuan and Jiao, Licheng and Liu, Fang and Sun, Qigong and Zhao, Jin , title =. Pattern Recognition , volume =. 2020 , type =

    2020

  25. [33]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =

    Liu, Mengkun and Jiao, Licheng and Liu, Xu and Li, Lingling and Liu, Fang and Yang, Shuyuan and Guo, Yuwei and Chen, Puhua , title =. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =. 2024 , type =

    2024

  26. [34]

    IEEE transactions on image processing , volume =

    Bi, Haixia and Xu, Lin and Cao, Xiangyong and Xue, Yong and Xu, Zongben , title =. IEEE transactions on image processing , volume =. 2020 , type =

    2020

  27. [35]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , ISSN =

    Wang, Lei and Zhu, Shenghui and Hong, Hanyu and Shi, Yu and Zhu, Ying and Ma, Lei , title =. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , ISSN =. 2025 , type =

    2025

  28. [36]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =

    Qin, Rui and Fu, Xiongjun and Lang, Ping , title =. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =. 2020 , type =

    2020

  29. [37]

    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =

    Liu, Hongying and Zhu, Tianwen and Shang, Fanhua and Liu, Yuanyuan and Lv, Derui and Yang, Shuyuan , title =. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , volume =. 2020 , type =

    2020

  30. [38]

    arXiv preprint arXiv:2104.00341 , year =

    Chakraborty, Tanmay and Trehan, Utkarsh , title =. arXiv preprint arXiv:2104.00341 , year =

    arXiv

  31. [39]

    International Journal of Applied Earth Observation and Geoinformation , volume =

    Alkhatib, Mohammed Q , title =. International Journal of Applied Earth Observation and Geoinformation , volume =. 2025 , type =

    2025

  32. [40]

    IEEE Transactions on Geoscience and Remote Sensing , ISSN =

    Shi, Junfei and Shen, Keyan and Jin, Haiyan and Zhang, Yuanlin and Hua, Wenqiang and Lv, Zhiyong and Gong, Maoguo and Lin, Weisi , title =. IEEE Transactions on Geoscience and Remote Sensing , ISSN =. 2025 , type =

    2025

Pith tools

Reviewed July 30, 2026 · model on record in the stance chip above.