Pith. sign in

REVIEW 3 major objections 7 minor 43 references

A Sensor Agnostic Domain Generalization Framework for Leveraging Geospatial Foundation Models: Enhancing Semantic Segmentation viaSynergistic Pseudo-Labeling and Generative Learning

T0 review · 3 major / 7 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read A 6-band multispectral foundation model adapts to 242-band hyperspectral and multispectral segmentation via entropy pseudo-labeling plus a source-to-target MAE whose gradient weights unlabeled target pixels by confidence.

desk verdict The empirical recipe works in their tests, but the advertised math for confidence-weighted MAE cancels to zero under their own equations; fix that and the water overclaim before trusting the paper. read the letter →

arxiv 2505.01558 v1 pith:L7WY3XAD submitted 2025-05-02 cs.CV

classification cs.CV
keywords domainadaptationsemanticsegmentationgeospatialfoundationmodelsmaskedautoencoderhyperspectralimagingpseudo-labelingremotesensingPrithvi
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper seeks to show that a geospatial foundation model pre-trained on six-band multispectral imagery can be repurposed for semantic segmentation in very different sensing regimes, including 242-band hyperspectral images and unseen geographic regions, without retraining its encoder. It proposes a multi-task loss that adds two auxiliary objectives to the supervised segmentation loss: an entropy-minimization term that soft-aligns pseudo-labels in the target domain, and a masked-autoencoder term that reconstructs masked target patches from a concatenation of source and unmasked target patch embeddings. The paper's central explanatory claim is mathematical: the MAE term re-weights each unlabeled target pixel's gradient contribution by the segmentation model's own confidence in that pixel, so confidently labeled regions drive adaptation more strongly. On the C2Seg-AB hyperspectral benchmark and the FLAIR multispectral benchmark, the full framework reports the best mean intersection-over-union among the compared methods (0.3835 and 0.5280), and ablations show the two auxiliary terms help most when used together.

What carries the argument

The load-bearing object is the total loss $L_{Tot} = L_{Seg} + \lambda_{DA} L_{DA} + \lambda_{MAE} L_{MAE}$ over one shared feature extractor $f_\theta$, with the segmentation head $h_{\theta_{seg}}$ and the MAE head $g_{\theta_M}$. The source-to-target MAE flow concatenates the source image's patch-embedding sequence with the unmasked target patch sequence along the sequence dimension, so a Conv1D blending layer has to recover the full target sequence from both domains; self-attention in the frozen ViT encoder then has to produce features that work across domains. The mathematical identity doing the explanatory work is $P(y_k \mid x^T_k, X_S; \bar{\theta}) = \mathbb{E}$ over the other target pixels of $h(f(Z, x^T_k), y_k)$, which converts the MAE gradient into a per-pixel, per-class confidence weight. Spectral adaptation is carried by a Conv2D layer that projects the input channels (e.g., 242 for HSI) into Prithvi's 6-band token space, followed by trainable adapter transformer layers.

What would settle it

In a controlled run on FLAIR, replace the per-pixel confidence weight $P(y_k \mid x^T_k, X_S; \bar{\theta})$ in the MAE gradient with a uniform weight while keeping everything else fixed; if mIoU does not change, the paper's proposed confidence-gating mechanism is not carrying the reported gain.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that MAE-based generative learning is not a generic regularizer but a confidence-gated weighting mechanism when its encoder is shared with the segmentation head. Writing the combined objective as a joint likelihood $J = P(Y_S, Y_T, X_T \mid X_S)$ and differentiating with respect to the shared encoder parameters $\theta$, the contribution of each unlabeled target pixel $k$ to the gradient is multiplied by $P(y_k \mid x^T_k, X_S; \bar{\theta})$, which equals the expected segmentation-model confidence in pixel $k$ over the other target pixels. Confident pixels therefore exert larger influence on the shared features, and uncertain pixels are suppressed; the entropy-alignment term $L_{DA}$ feeds this mechanism by pushing target predictions toward confidence in the first place. The paper further claims that this framework is the first to extend MAE-based reconstruction of a multispectral foundation model to hyperspectral inputs, using a Conv2D spectral-adaptation layer and added adapter transformer blocks around Prithvi's frozen encoder.

Load-bearing premise

The framework assumes that Prithvi's frozen encoder, pretrained on 6-channel multispectral imagery, still provides useful semantic features for a 242-band hyperspectral image after a single Conv2D projection and adapter tuning; if the frozen representations do not transfer across that spectral gap, the projection and adapters have no signal to recover and the whole source-to-target MAE and segmentation pipeline reduces to noise.

Editorial extensions

If this is right

  • If the central claim holds, a single foundation model pretrained on a widely available multispectral sensor can serve as the backbone for segmentation across multispectral, hyperspectral, and other optical sensors, with only a small channel-projection layer and adapters retrained.
  • The MAE objective's effect is conditional on the segmentation head's confidence: it accelerates learning on target pixels the model already handles and slows learning where it is uncertain, so its value cannot be assessed in isolation from the alignment loss.
  • Because source and target patch sequences are concatenated in embedding space rather than pixel space, the framework can pair domains with different spatial resolutions in the same training batch.
  • The reported C2Seg-AB and FLAIR results imply the method also copes with open-set domain shifts, such as target classes that never appear in the source labels.

Reading between the lines

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

  • A natural testable extension is to make the confidence weights explicit and compare them against the implicit MAE weighting, e.g., by replacing $P(y_k \mid x^T_k, X_S)$ with a thresholded or temperature-scaled version; matching behavior would confirm the mechanism and allow direct control of it.
  • The mathematical claim suggests the auxiliary MAE loss will inherit any calibration error of the segmentation head: a miscalibrated model that is overconfident on wrong target pixels would amplify bad gradients, which may explain the ablation result where $L_{Seg}+L_{MAE}$ alone degrades performance on C2Seg-AB.
  • The framework's sensor-agnostic claim rests on the Conv2D projection bridging a 6-band to 242-band gap; an isolated probe of that bridge, such as reconstructing HSI patches from projected frozen features with no adapters, would separate Prithvi's contribution from the adapters' contribution.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper proposes a domain-adaptation framework that adapts the Prithvi geospatial foundation model to new sensors by keeping its encoder frozen and adding a Conv2D spectral-projection layer plus trainable transformer adapters, a segmentation head, and an MAE reconstruction head. The total loss (Eq. 1) combines supervised cross-entropy over the labeled source and a small set of labeled target pixels (LSeg), entropy-minimization soft pseudo-label alignment (LDA), and a masked-autoencoder loss in which a token sequence concatenating all source patches with unmasked target patches is used to reconstruct masked target patches (LMAE). Experiments on C2Seg-AB (242-band HSI, Berlin to Augsburg) and FLAIR (multispectral, France) report the highest mIoU among the compared methods (0.3835 and 0.5280), with ablations and a claimed rigorous mathematical analysis (contribution 3) intended to show that the MAE objective acts on unlabeled target pixels through a confidence-based dynamic weighting mechanism tied to the segmentation model. The supplementary material contains the derivation, additional ablations on C2Seg-AB, and spatial and spectral reconstruction visualizations.

Significance. If the empirical results are taken at face value, the framework is a practically valuable contribution: it achieves the best reported mIoU on two benchmarks across distinct sensing regimes, is evaluated with 10 random seeds, uses official baseline implementations for most comparisons, and the authors release the code. The FLAIR open-set scenario and the spatial and spectral reconstruction evaluations are useful and go beyond the usual single-benchmark DA studies. However, the paper's advertised theoretical contribution does not survive scrutiny: the derivation in Supp. §A.2 cancels the terms on which the dynamic weighting claim is based, so contribution 3 and the mechanism explanation of the ablations are unsupported and must be revised or removed. The empirical claim of universal superiority is also statistically fragile on C2Seg-AB, where the margin over the closest baseline is within one standard error. With the mathematical claim corrected or retracted, significance tests added, and the identified ablations performed, the empirical framework could form a solid paper.

major comments (3)
  1. [Supp. §A.2, Eqs. (21)-(27); main text §3.4, Eq. (9)] The claimed dynamic-weighting mechanism does not follow from the derivation and, under the paper's own factorization, cancels identically. Substituting the joint model P(xT_k, yk | XS; θ̄) = P(yk | xT_k, XS; θ̄) · N(xT_k; g(f(XS,k)), Σ) (Eq. 25) into the identity (22) gives Σ_yk P(yk|x)·∂θ log P(yk|x) + Σ_yk P(yk|x)·∂θ log N(xT_k; ·). The first sum is zero at every parameter value because Σ_yk P(yk | xT_k, XS; θ̄) = 1 (the segmentation head outputs a normalized softmax), so ∂θ Σ_yk P(yk|x) = 0. The third term of Eq. (23) therefore equals the ordinary gradient of the marginal reconstruction log-likelihood, in which hθseg does not appear; Eq. (9) in the main text inherits this problem, and the regulation described around Eq. (27) is vacuous because the weights multiply terms that cancel exactly. In addition, Eq. (21) writes ∂J/∂θ on the left while the right-hand side differentiates log J, a factor-of-J error, although the cancellation is independent of it. Because contribution 3 and the mechanism explanation in §4.3 rest entirely on this derivation, the authors should either retract the claim or introduce confidence weighting directly into the objective (e.g., a segmentation-confidence-weighted MAE loss) and validate it empirically.
  2. [Tables 1 and 3] The text claims the method significantly outperforms all compared methods, but no significance testing is reported, and on C2Seg-AB the headline margin over the closest competitor is within noise: Ours (mIoU 0.3835 ± 0.0161) versus PCS (0.3731 ± 0.0148), difference approximately 0.0104, pooled standard error across the 10 runs approximately 0.0069, t ≈ 1.5, p ≈ 0.15. On FLAIR the margin over MIC (0.5280 ± 0.0223 versus 0.4938 ± 0.0350, t ≈ 2.6) does appear significant, so the situation differs across datasets. Report paired tests over the 10 seeds for every headline comparison (Ours versus each baseline on both datasets, for MA, mIoU, and mF1) and temper the universal claims in §4.2 accordingly.
  3. [§4.1 (Network Architecture) and §4.3] The paper never isolates whether the reported gains come from the frozen pretrained representations, from the Conv2D projection and adapters, or from the proposed losses. The zero-shot column and all DA baselines use the same Prithvi backbone, so a reader cannot tell whether the framework is leveraging the foundation model or merely benefiting from the domain-alignment and generative objectives; the current evidence is only indirect (final metrics and MAE reconstruction quality). To substantiate contribution 1 and the sensor-agnostic claim, add controls: (i) the identical architecture with Prithvi's weights replaced by random initialization, trained on the same losses; (ii) full fine-tuning of Prithvi's encoder instead of the frozen-plus-adapter scheme; and (iii) a sensitivity check on the spectral-projection layer (the 242-to-6 channel mapping that is the load-bearing bridge for the hyperspectral experiment).
minor comments (7)
  1. [Supp. §E] The statement that adding LMAE alone degrades performance similar to Table 3 is contradicted by Table 3, where LSeg+LMAE (mIoU 0.5156) improves over LSeg (0.3239); the claim holds only for the C2Seg-AB ablation (Table 4). The sign of LMAE's marginal effect is dataset-dependent, so the paragraph should be rewritten and the synergy explanation qualified accordingly.
  2. [§4.2, FLAIR paragraph] The claim that for the two open-set classes (plowed land and water) the method achieves the best performance is inaccurate for water, where Table 2 reports Ours at 0.0421 versus CIA-UDA at 0.0794 for F1.
  3. [§3.2, §4.1, Supp. §B] The MAE masking ratio used during joint training of the full framework and the Conv1D parameters K' and S' are never specified; Supp. §D states only the 50 percent ratio used for the standalone generative evaluation. Please report these values for reproducibility.
  4. [Eq. (7)] Shanon Entropy should read Shannon entropy.
  5. [Supp. §A.1] Model Defentions should be Model Definitions, and An observation of of the random variable contains a doubled of.
  6. [Table 3 caption and §4.3] The second column is called the baseline in the text but is not labeled as such in the table, and the Zero Shot column is a source-only fine-tuned model rather than a zero-shot deployment; please rename the columns to avoid confusion.
  7. [Figure 3] The parameter boxes in Figure 3 are barely legible in the rendered manuscript, with kernel sizes and filter counts appearing blank, which undermines the reproducibility of the architecture description.

Circularity Check

1 steps flagged · score 6.0 of 10

Contribution 3's 'dynamic weighting' insight collapses under the paper's own factorization: the hθseg-dependent terms cancel, leaving the ordinary MAE gradient.

  1. other [Supplementary Sec. A.2.2, Eqs. (23) and (25); main text Eq. (9), Sec. 3.4]
    "P(yk|xTk, XS, ¯θ)×N (xTk ;g(f(XS,k )), Σ) (25) ... Applying the trick in Equation (22) for the last term in Equation (21), we got : ... Σyk P(yk|xTk, XS; ¯θ) × ∂∂θ log P(xTk,yk| XS; ¯θ) (23)"

    Under Eq. (25), the joint in the last term of Eq. (23) factorizes as P(y,x|XS)=P(y|x;θ,θseg)·N(x;g(f(XS)),Σ) with P(y|x)=h(f(XT),y). Substituting into the y-sum gives Σ_y h(y|x)[∂log h(y|x)/∂θ + ∂log N/∂θ]; the h-dependent piece is Σ_y ∂h(y|x)/∂θ = ∂/∂θ Σ_y h(y|x) = 0 because the softmax head normalizes for every θ. The surviving term is the ordinary MAE gradient ∂/∂θ log N(x;g(f(XS)),Σ). Hence the claimed dynamic weighting ϕ1 in Eq. (9) cancels identically; the result reduces, by the paper's own equations, to the plain MAE term already present in the input LMAE. The implemented loss in Eqs. (4)-(5) contains no class variable, so the confidence-regulated mechanism is not a derived property but an artifact of the chosen factorization.

full rationale

The empirical framework is tested against external baselines (GDA, PCS, CDS, MIC, CIAUDA, UDAMEBS) and ablations on two datasets, so those results do not reduce to any fitted parameter or self-citation. No load-bearing self-citation chain is present; Prithvi, ADVENT, and Lasserre are external. The circularity is confined to contribution 3: the mathematical 'insight' that MAE weights unlabeled target pixels by segmentation confidence is obtained by writing the unlabeled term with the Lasserre identity and defining the artificial joint P(x,y)=P(y|x)N(x); under that same definition the y-sum cancels the h-dependence, leaving the ordinary MAE gradient. The claimed first-principles result is therefore equivalent, by construction, to the input LMAE term, not an independent derivation from it. This is a partial circularity and overclaim in one of the paper's three headline contributions; the central empirical claim retains independent content.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The method relies on the frozen Prithvi backbone and the transferability assumption, two tuning weights set to 1, and a handful of architectural choices. No new physical entities are introduced.

free parameters (3)
  • λDA and λMAE = 1
    Set empirically to 1 in all experiments; affects the balance between segmentation, entropy minimization, and reconstruction.
  • Conv1D layer parameters K' and S' = chosen to match target sequence length
    The kernel and stride of the sequence-blending Conv1D are selected to recover the target patch count, a design choice not justified by a principle.
  • MAE masking ratio during training = not reported for training
    The paper evaluates masking ratios of 50, 75, and 100 percent in the generative analysis but does not state the training-time masking ratio, which is a free design parameter.
assumptions (4)
  • domain assumption Prithvi's frozen 6-channel multispectral features transfer to other spectral modalities via a linear projection
    Section 3 Network Architecture assumes a single Conv2D layer can map HSI (242 bands) and VHR RGB onto Prithvi's input space while preserving semantic content; the whole method depends on this.
  • domain assumption Conditional independence of YS, YT, XT given XS and model parameters
    Supp. Eq. (15)-(16) are used to factor the joint likelihood; this independence is not verified for satellite imagery.
  • domain assumption Gaussian noise model for MAE reconstruction
    Supp. Eq. (11) justifies the MSE loss as a Gaussian log-likelihood, which is standard for MAE but not scrutinized.
  • standard math The 'trick' from [19] for rewriting ∂log P(x)/∂θ
    Supp. Eq. (22) is a standard probability identity, properly attributed to [19].

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Sensor Agnostic Domain Generalization Framework for Leveraging Geospatial Foundation Models: Enhancing Semantic Segmentation viaSynergistic Pseudo-Labeling and Generative Learning." pith.science (2026). https://pith.science/paper/L7WY3XAD

@misc{pith2026250501558,
  author       = {Pith},
  title        = {Pith review of: A Sensor Agnostic Domain Generalization Framework for Leveraging Geospatial Foundation Models: Enhancing Semantic Segmentation viaSynergistic Pseudo-Labeling and Generative Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L7WY3XAD}},
  note         = {Machine review of arXiv:2505.01558}
}
read the original abstract

Remote sensing enables a wide range of critical applications such as land cover and land use mapping, crop yield prediction, and environmental monitoring. Advances in satellite technology have expanded remote sensing datasets, yet high-performance segmentation models remain dependent on extensive labeled data, challenged by annotation scarcity and variability across sensors, illumination, and geography. Domain adaptation offers a promising solution to improve model generalization. This paper introduces a domain generalization approach to leveraging emerging geospatial foundation models by combining soft-alignment pseudo-labeling with source-to-target generative pre-training. We further provide new mathematical insights into MAE-based generative learning for domain-invariant feature learning. Experiments with hyperspectral and multispectral remote sensing datasets confirm our method's effectiveness in enhancing adaptability and segmentation.

Figures

Figures reproduced from arXiv: 2505.01558 by the authors.

Figure 1
Figure 1. Illustration of the proposed multi-task learning frame [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the target and source domains used from [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Prithvi model with adapters: feature extractor [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Comparative Segmentation Inference using the C2Seg-AB Dataset. Display includes (a) Ground Truth Mask, (b) Our Method, [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Comparative Segmentation Inference using FLAIR Dataset. Display includes (a) Ground Truth Mask, (b) Our Method, (c) PCS, [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]
Figure 6
Figure 6. Figure 6: Generative task evaluation on HSI modality with C2Seg-AB dataset, showing original, masked, and reconstructed images across [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]
Figure 7
Figure 7. Figure 7: Mean and standard deviation of the spectral distribution for the remaining classes not covered in the main text, evaluated for [PITH_FULL_IMAGE:figures/full_fig_p017_7.png]
Figure 8
Figure 8. Figure 8: Generative task evaluation on MSI modality with FLAIR dataset, showing original, masked, and reconstructed images across [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 36 canonical work pages

  1. [1]

    ExPLoRA: Parameter-efficient extended pre- training to adapt vision transformers under domain shifts

    Anonymous. ExPLoRA: Parameter-efficient extended pre- training to adapt vision transformers under domain shifts. In Submitted to The Thirteenth International Conference on Learning Representations, 2024. 1

  2. [2]

    A partially unsupervised cascade classifier for the analysis of multitem- poral remote-sensing images

    Lorenzo Bruzzone and Diego Fern ´andez Prieto. A partially unsupervised cascade classifier for the analysis of multitem- poral remote-sensing images. Pattern Recognition Letters, 23(9):1063–1071, 2002. 1

  3. [3]

    Unsupervised domain adaptation for semantic seg- mentation of high-resolution remote sensing imagery driven by category-certainty attention

    Jie Chen, Jingru Zhu, Ya Guo, Geng Sun, Yi Zhang, and Min Deng. Unsupervised domain adaptation for semantic seg- mentation of high-resolution remote sensing imagery driven by category-certainty attention. IEEE Transactions on Geo- science and Remote Sensing, 60:1–15, 2022. 7

  4. [4]

    In Proceedings of the 37th International Conference on Machine Learning , pages 1691–1703

    Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Hee- woo Jun, David Luan, and Ilya Sutskever. In Proceedings of the 37th International Conference on Machine Learning , pages 1691–1703. PMLR, 2020. 2, 3

  5. [5]

    Semantic segmentation of remote sensing images using transfer learning and deep convolutional neural network with dense connection

    Binge Cui, Xin Chen, and Yan Lu. Semantic segmentation of remote sensing images using transfer learning and deep convolutional neural network with dense connection. IEEE Access, 8:116744–116755, 2020. 1

  6. [6]

    On the generalization ability of unsupervised pre- training

    Yuyang Deng, Junyuan Hong, Jiayu Zhou, and Mehrdad Mahdavi. On the generalization ability of unsupervised pre- training. In Proceedings of The 27th International Confer- ence on Artificial Intelligence and Statistics , pages 4519–

  7. [7]

    An image is worth 16 x 16 words :

    A Dosovitskiy, L Beyer, A Kolesnikov, D Weissenborn, X Zhai, T Unterthiner, M Dehghani, M Minderer, G Heigold, S Gelly, J Uszkoreit, and N Houslby. An image is worth 16 x 16 words :. International Conference on Learning Repre- sentations, pages 1–21, 2021. 2

  8. [8]

    FLAIR #1: semantic seg- mentation and domain adaptation dataset

    Anatol Garioud, St ´ephane Peillet, Eva Bookjans, S ´ebastien Giordano, and Boris Wattrelos. FLAIR #1: semantic seg- mentation and domain adaptation dataset. pages 1–9, 2022. 5

Show all 43 references
  1. [9]

    Flair : a country-scale land cover se- mantic segmentation dataset from multi-source optical im- agery

    Anatol Garioud, Nicolas Gonthier, Loic Landrieu, Apolline De Wit, Marion Valette, Marc Poup ´ee, Sebastien Giordano, and boris Wattrelos. Flair : a country-scale land cover se- mantic segmentation dataset from multi-source optical im- agery. In Advances in Neural Information P...

  2. [10]

    Copernicus sentinel-2a cali- bration and products validation status

    Ferran Gascon, Catherine Bouzinac, Olivier Th ´epaut, Math- ieu Jung, Benjamin Francesconi, J ´erˆome Louis, Vincent Lonjou, Bruno Lafrance, St ´ephane Massera, Ang ´elique Gaudel-Vacaresse, Florie Languille, Bahjat Alhammoud, Franc ¸oise Viallefont, Bringfried Pflug, Jakub Bi...

  3. [11]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll´ar, and Ross Girshick. Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 16000–16009, 2022. 2

  4. [12]

    Cross-city matters: A multi- modal remote sensing benchmark dataset for cross-city se- mantic segmentation using high-resolution domain adapta- tion networks

    Danfeng Hong, Bing Zhang, Hao Li, Yuxuan Li, Jing Yao, Chenyu Li, Martin Werner, Jocelyn Chanussot, Alexander Zipf, and Xiao Xiang Zhu. Cross-city matters: A multi- modal remote sensing benchmark dataset for cross-city se- mantic segmentation using high-resolution domain adapt...

  5. [13]

    Spectralgpt: Spectral remote sensing foundation model

    Danfeng Hong, Bing Zhang, Xuyang Li, Yuxuan Li, Chenyu Li, Jing Yao, Naoto Yokoya, Hao Li, Pedram Ghamisi, Xiup- ing Jia, Antonio Plaza, Paolo Gamba, Jon Atli Benediktsson, and Jocelyn Chanussot. Spectralgpt: Spectral remote sensing foundation model. IEEE Transactions on Patte...

  6. [14]

    Mic: Masked image consistency for context- enhanced domain adaptation

    Lukas Hoyer, Dengxin Dai, Haoran Wang, and Luc Van Gool. Mic: Masked image consistency for context- enhanced domain adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11721–11732, 2023. 6, 8

  7. [15]

    Semi-supervised bidirectional alignment for remote sensing cross-domain scene classification

    Wei Huang, Yilei Shi, Zhitong Xiong, Qi Wang, and Xiao Xiang Zhu. Semi-supervised bidirectional alignment for remote sensing cross-domain scene classification. ISPRS Journal of Photogrammetry and Remote Sensing , 195:192– 203, 2023. 1

  8. [16]

    Foundation models for generalist geospatial artificial intelligence, 2023

    Johannes Jakubik, S Roy, CE Phillips, P Fraccaro, D Godwin, B Zadrozny, D Szwarcman, C Gomes, G Nyir- jesy, B Edwards, et al. Foundation models for generalist geospatial artificial intelligence, 2023. URL https://arxiv. org/abs/2310.18660. 2, 3

  9. [17]

    Matthew Davis, David B

    Neal Jean, Marshall Burke, Michael Xie, W. Matthew Davis, David B. Lobell, and Stefano Ermon. Combining satellite imagery and machine learning to predict poverty. Science, 353(6301):790–794, 2016. 1

  10. [18]

    Plummer, Stan Sclaroff, and Kate Saenko

    Donghyun Kim, Kuniaki Saito, Tae-Hyun Oh, Bryan A. Plummer, Stan Sclaroff, and Kate Saenko. Cds: Cross- domain self-supervised pre-training. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 9123–9132, 2021. 3, 6, 8

  11. [19]

    Hybrids of generative and discrimi- native methods for machine learning

    Julia Aur ´elie Lasserre. Hybrids of generative and discrimi- native methods for machine learning. PhD thesis, University of Cambridge, 2008. 3, 2

  12. [20]

    Brooks, Fahim Tajwar, Marshall Burke, Stefano Ermon, David B

    Jihyeon Lee, Nina R. Brooks, Fahim Tajwar, Marshall Burke, Stefano Ermon, David B. Lobell, Debashish Biswas, and Stephen P. Luby. Scalable deep learning to identify brick kilns and aid regulatory capacity. Proceedings of the Na- tional Academy of Sciences, 118(17):e2018863118, 2021. 1

  13. [21]

    Cost-sensitive self-paced learning 9 with adaptive regularization for classification of image time series

    Hao Li, Jianzhao Li, Yue Zhao, Maoguo Gong, Yujing Zhang, and Tongfei Liu. Cost-sensitive self-paced learning 9 with adaptive regularization for classification of image time series. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 14:11713–117...

  14. [22]

    Vision-language models in remote sensing: Current progress and future trends

    Xiang Li, Congcong Wen, Yuan Hu, Zhenghang Yuan, and Xiao Xiang Zhu. Vision-language models in remote sensing: Current progress and future trends. IEEE Geoscience and Remote Sensing Magazine, 12(2):32–66, 2024. 2

  15. [23]

    Unsupervised domain adaptation semantic segmentation of remote sensing images with mask enhancement and balanced sampling

    Xin Li, Yuanbo Qiu, Jixiu Liao, Fan Meng, and Peng Ren. Unsupervised domain adaptation semantic segmentation of remote sensing images with mask enhancement and balanced sampling. IEEE Transactions on Geoscience and Remote Sensing, pages 1–1, 2025. 6, 7, 8

  16. [24]

    Diverse hyperspectral remote sensing image synthesis with diffusion models

    Liqin Liu, Bowen Chen, Hao Chen, Zhengxia Zou, and Zhenwei Shi. Diverse hyperspectral remote sensing image synthesis with diffusion models. IEEE Transactions on Geo- science and Remote Sensing, 61:1–16, 2023. 2

  17. [25]

    Decoupled Weight De- cay Regularization

    Ilya Loshchilov and Frank Hutter. Decoupled Weight De- cay Regularization. In International Conference on Learning Representations, 2019. 3

  18. [26]

    Domain-adversarial training of self- attention-based networks for land cover classification using multi-temporal sentinel-2 satellite imagery

    Mauro Martini, Vittorio Mazzia, Aleem Khaliq, and Mar- cello Chiaberge. Domain-adversarial training of self- attention-based networks for land cover classification using multi-temporal sentinel-2 satellite imagery. Remote Sensing, 13(13), 2021. 1

  19. [27]

    Foundation models for generalist medi- cal artificial intelligence

    Michael Moor, Oishi Banerjee, Zahra Shakeri Hossein Abad, Harlan M Krumholz, Jure Leskovec, Eric J Topol, and Pranav Rajpurkar. Foundation models for generalist medi- cal artificial intelligence. Nature, 616(7956):259–265, 2023. 2

  20. [28]

    Category-level assignment for cross-domain semantic segmentation in remote sensing images

    Huan Ni, Qingshan Liu, Haiyan Guan, Hong Tang, and Joce- lyn Chanussot. Category-level assignment for cross-domain semantic segmentation in remote sensing images. IEEE Transactions on Geoscience and Remote Sensing , 61:1–16,

  21. [29]

    Domain adaptation in remote sensing image classification: A survey

    Jiangtao Peng, Yi Huang, Weiwei Sun, Na Chen, Yujie Ning, and Qian Du. Domain adaptation in remote sensing image classification: A survey. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 15:9842– 9859, 2022. 1

  22. [30]

    Tensor alignment based domain adaptation for hyperspectral image classifica- tion

    Yao Qin, Lorenzo Bruzzone, and Biao Li. Tensor alignment based domain adaptation for hyperspectral image classifica- tion. IEEE Transactions on Geoscience and Remote Sensing, 57(11):9290–9307, 2019. 1

  23. [31]

    Parameter efficient self-supervised geospatial domain adap- tation

    Linus Scheibenreif, Michael Mommert, and Damian Borth. Parameter efficient self-supervised geospatial domain adap- tation. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 27841– 27851, 2024. 1, 3, 6, 8

  24. [32]

    Transfer sparse subspace analysis for unsupervised cross-view scene model adaptation

    Hao Sun, Shuai Liu, Shilin Zhou, and Huanxin Zou. Transfer sparse subspace analysis for unsupervised cross-view scene model adaptation. IEEE Journal of Selected Topics in Ap- plied Earth Observations and Remote Sensing , 9(7):2901– 2909, 2016. 1

  25. [33]

    Ringmo: A remote sensing foundation model with masked image modeling

    Xian Sun, Peijin Wang, Wanxuan Lu, Zicong Zhu, Xiao- nan Lu, Qibin He, Junxi Li, Xuee Rong, Zhujun Yang, Hao Chang, Qinglin He, Guang Yang, Ruiping Wang, Jiwen Lu, and Kun Fu. Ringmo: A remote sensing foundation model with masked image modeling. IEEE Transactions on Geo- scien...

  26. [34]

    Llama: open and efficient foundation language models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timoth ´ee Lacroix, Bap- tiste Rozi `ere, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: open and efficient foundation language models. arxiv. arXiv preprint arXiv:2302.13971, 2023. 2

  27. [35]

    Advent: Adversarial entropy min- imization for domain adaptation in semantic segmentation

    Tuan-Hung Vu, Himalaya Jain, Maxime Bucher, Matthieu Cord, and Patrick Perez. Advent: Adversarial entropy min- imization for domain adaptation in semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019. 1, 2, 3, 4

  28. [36]

    Fifty years of landsat science and impacts.Remote Sensing of Environment, 280:113195, 2022

    Michael A Wulder, David P Roy, V olker C Radeloff, Thomas R Loveland, Martha C Anderson, David M Johnson, Sean Healey, Zhe Zhu, Theodore A Scambos, Nima Pahle- van, et al. Fifty years of landsat science and impacts.Remote Sensing of Environment, 280:113195, 2022. 1

  29. [37]

    Parameter-efficient fine-tuning for pre-trained vision models: A survey

    Yi Xin, Siqi Luo, Haodi Zhou, Junlong Du, Xiaohong Liu, Yue Fan, Qing Li, and Yuntao Du. Parameter-efficient fine-tuning for pre-trained vision models: A survey. arXiv preprint arXiv:2402.02242, 2024. 2, 4, 6

  30. [38]

    Un- derstanding how pretraining regularizes deep learning algo- rithms

    Yu Yao, Baosheng Yu, Chen Gong, and Tongliang Liu. Un- derstanding how pretraining regularizes deep learning algo- rithms. IEEE Transactions on Neural Networks and Learn- ing Systems, 34(9):5828–5840, 2023. 3

  31. [39]

    Deep gaussian process for crop yield predic- tion based on remote sensing data

    Jiaxuan You, Xiaocheng Li, Melvin Low, David Lobell, and Stefano Ermon. Deep gaussian process for crop yield predic- tion based on remote sensing data. Proceedings of the AAAI Conference on Artificial Intelligence, 31(1), 2017. 1

  32. [40]

    Metaearth: A generative foundation model for global-scale remote sensing image generation

    Zhiping Yu, Chenyang Liu, Liqin Liu, Zhenwei Shi, and Zhengxia Zou. Metaearth: A generative foundation model for global-scale remote sensing image generation. arXiv preprint arXiv:2405.13570, 2024. 2, 3

  33. [41]

    Florence: A new foundation model for computer vision

    Lu Yuan, Dongdong Chen, Yi-Ling Chen, Noel Codella, Xiyang Dai, Jianfeng Gao, Houdong Hu, Xuedong Huang, Boxin Li, Chunyuan Li, et al. Florence: A new foundation model for computer vision. arXiv preprint arXiv:2111.11432, 2021. 2

  34. [42]

    Prototypical cross-domain self-supervised learn- ing for few-shot unsupervised domain adaptation

    Xiangyu Yue, Zangwei Zheng, Shanghang Zhang, Yang Gao, Trevor Darrell, Kurt Keutzer, and Alberto Sangiovanni Vincentelli. Prototypical cross-domain self-supervised learn- ing for few-shot unsupervised domain adaptation. In Pro- ceedings of the IEEE/CVF Conference on Computer V...

  35. [43]

    Remote sensing im- age transfer classification based on weighted extreme learn- ing machine

    Yang Zhou, Jie Lian, and Min Han. Remote sensing im- age transfer classification based on weighted extreme learn- ing machine. IEEE Geoscience and Remote Sensing Letters, 13(10):1405–1409, 2016. 1 10 A Sensor Agnostic Domain Generalization Framework for Leveraging Geospatial F...

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.