Pith. sign in

REVIEW 4 major objections 6 minor 55 references

A Transfer Learning-Based Method for Water Body Segmentation in Remote Sensing Imagery: A Case Study of the Zhada Tulin Area

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Fine-tuning a SegFormer pre-trained on diverse water imagery lifts water-body segmentation overlap in the arid Zhada Tulin area from 25.50% to 64.84%.

desk verdict Overlapping crops and a patch-level split undermine the headline IoU; otherwise an honest but incremental transfer-learning case study. read the letter →

arxiv 2507.10084 v2 pith:BCPYKRUZ submitted 2025-07-14 cs.CV cs.LG

Haonan Chen (Tibet University) , Xin Tong (Northwestern Polytechnical University) This is my paper · ORCID
classification cs.CVcs.LG
keywords remotesensingimagerywaterbodysegmentationtransferlearningSegFormerdomainshiftZhadaTulinsemanticTibetanPlateau
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that a standard two-step recipe—pre-train a segmentation model on a large, diverse collection of water imagery, then fine-tune it on a small labeled set from the place you actually care about—is enough to overcome both domain shift and data scarcity in arid high-mountain water monitoring. Using the SegFormer architecture, the study reports that fine-tuning on 162 patches from the Zhada Tulin area of Tibet lifts the water-body Intersection over Union (IoU) on an 18-patch validation set from 25.50% (direct transfer) to 64.84%, while scratch-trained baselines sit between 37.47% and 48.82%. If the result is right, high-resolution water mapping becomes practical in data-poor, climate-sensitive headwater basins without huge labeling campaigns. The resulting map also implies a concrete hydrological pattern: more than 80% of the water surface area is confined to less than 20% of the river channel length, a concentration relevant to water management and flash-flood risk.

What carries the argument

The load-bearing mechanism is the two-stage transfer pipeline built on SegFormer, a Transformer-based semantic segmentation network with a hierarchical multi-scale Transformer encoder and a lightweight all-MLP decoder. In stage one, the encoder starts from large-scale natural-image pretrained weights and the full model is trained on 3,875 high-resolution patches covering diverse plateau, lake, and mountain water bodies; in stage two, all weights are initialized from that source model and fine-tuned on 180 patches from the arid target site. The training objective is a compound loss of weighted binary cross-entropy and Dice loss, with the rare water class weighted at 0.7711 against 0.2289 for background. The authors deliberately keep the raw radiometric differences between the two satellite sensors in the data, so the learned features have to be robust to cross-sensor spectral variation.

What would settle it

Label a held-out set of full Gaofen-2 scenes over the Zhada Tulin area that were not used in the 90:10 split, run the fine-tuned A2 model over entire scenes, and compare the resulting water overlap score with the 64.84% patch-level number; if the full-scene score falls back toward the 25.50% direct-transfer level, the reported gain is an artifact of patch selection and the 80/20 concentration statistic would need re-estimation.

Watch

Extended reading notes

Core claim

The central claim is that a two-stage transfer-learning strategy, rather than a better network or more data, is what produces high-precision water-body segmentation in a domain-shifted small-sample setting. The paper reports that fine-tuning a source-domain SegFormer model on the Zhada Tulin target raises water IoU from 25.50% under direct transfer to 64.84%, and that this level of detail is enough to reveal an 80/20 concentration of water area along the Xiangquan river corridor. On the paper's account, the source-domain pre-training supplies robust low-level feature extraction that survives sensor and landscape differences, while target fine-tuning adapts high-level semantics to the target's turbid, gully-confined water and sediment-heavy background. The authors take the result as evidence that the 'general-to-specific' knowledge transfer is the operative mechanism, and that the fine-tuned mask is trustworthy enough to support geoscientific statements about tectonic control and corridor effects in the drainage system.

Load-bearing premise

The load-bearing assumption is that the single random 90:10 split of 180 target patches, with 162 training and 18 validation patches, produces a representative validation IoU; the paper reports no variance across splits or seeds, so a different split could shift the headline 25.50%-to-64.84% comparison.

Editorial extensions

If this is right

  • If the 64.84% IoU figure is representative, the same pre-train-then-fine-tune recipe can be applied to other data-scarce arid basins with only a few hundred labeled patches, reusing the same source model.
  • At the reported accuracy, the water mask is detailed enough to serve as input for hydrological analyses such as measuring channel length and water-area concentration, so the 80/20 statistic becomes a testable geoscientific claim rather than a visual impression.
  • The comparison against scratch-trained SegFormer and U-Net baselines implies that the gain comes mainly from the transfer step, so reporting a model's architecture alone is not enough in domain-shifted settings.
  • A multi-temporal extension of the fine-tuned model, identified in the paper as future work, would convert the static concentrated-water map into a monitoring baseline for earlier snowmelt runoff and flash-flood-prone corridor zones.

Reading between the lines

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

  • A testable extension: fine-tune the same source model on a second arid site with a similarly small label budget and measure the IoU drop; a small drop would generalize the recipe, while a large drop would show the reported result is site-specific.
  • Because only patches containing water were retained, an editorially inferred question is whether full-scene deployment, with mostly dry terrain, keeps the same IoU or becomes dominated by false positives on dry sediment.
  • Since the paper keeps sensor radiometric differences uncorrected, the method could plausibly be applied to mixed-sensor satellite archives without normalization; if valid, historical imagery from multiple satellites could be pooled to build longer water records for the plateau.
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

4 major / 6 minor

Summary. The paper proposes a two-stage transfer learning strategy for water body segmentation in remote sensing imagery, targeting the arid Zhada Tulin area of the Tibetan Plateau. A SegFormer model (MiT-B5) is first pretrained on a diverse source domain of GF-7 imagery, then fine-tuned on a small target dataset of GF-2 imagery. The authors report that fine-tuning lifts water-body IoU from 25.50% (direct transfer) to 64.84%, outperforming scratch-trained SegFormer and U-Net baselines. They also report a geoscientific finding that over 80% of the water area concentrates in less than 20% of the river channel length, interpreted as a 'corridor effect.' The experimental design uses sliding-window cropping with 128-pixel stride on 512x512 patches, random 9:1 train/validation splitting, and compound BCE+Dice loss. The central claims rest on a single split with 18 validation patches per class.

Significance. If validated with a leakage-free experimental protocol, the proposed recipe would provide concrete evidence that fine-tuning a transformer-based segmentation model pretrained on diverse water imagery can substantially mitigate both domain shift and small-sample scarcity, a common barrier in climate-sensitive remote sensing. The paper also attempts to translate technical segmentation gains into a hydrologically meaningful statement about water concentration in this arid plateau region. However, the evaluation methodology as described does not currently support the headline quantitative claims because the training and validation patches are derived from heavily overlapping windows without spatial separation, and because the geoscientific concentration claim is stated without any reproducible measurement procedure.

major comments (4)
  1. [Section 2.2 (Sliding Window Cropping) and Section 3.2 (Table 2)] The sliding-window protocol uses a 512x512 window with a stride of 128 pixels, producing patches that overlap by 75% of their area. Since the 180 target-domain patches are then randomly split 9:1 into training and validation without image-level separation, a validation patch will typically be nearly identical to a training patch from the same GF-2 image, differing only by a 128-pixel shift. This means the fine-tuned model A2 may have memorized the validation content during training, while the directly transferred model A1 has not been exposed to any target training patches. Consequently, the headline comparison 25.50% (direct transfer) vs. 64.84% (fine-tuned) conflates spatial memorization with genuine transfer learning. To support the central claim, the authors must re-evaluate using spatially disjoint patches (e.g., non-overlapping windows) or, preferably, hold out entire images from training.
  2. [Table 2 (Model Performance Evaluation)] All validation metrics are reported on a single random split producing only 18 validation patches, with no standard deviation across splits or training seeds. Given the small and spatially autocorrelated validation set, the headline IoU numbers (25.50%, 37.47%, 48.82%, 64.84%) are unlikely to be statistically stable; a different split could materially change the ordering or magnitude. The authors should report mean and standard deviation over multiple random splits and training seeds, and ideally provide per-patch IoU distributions or a significance test for the comparison between A2 and the baselines.
  3. [Section 4.1 (Geoscientific Implications)] The statement that 'over 80% of the water surface area is confined to less than 20% of the total river channel length' is a key quantitative finding of the paper, yet the manuscript provides no methodological description of how 'river channel length' is measured, whether the statistic is derived from the segmentation map or from ancillary vector data, or what threshold/algorithm yields the 80/20 statement. Without this information, the result is not reproducible, and it is not possible to assess its sensitivity to segmentation errors. The authors need to define the measurement procedure, present the underlying calculation, and ideally test robustness against varying IoU thresholds.
  4. [Section 3.2 (B-scratch (U-Net) baseline)] The U-Net baseline is only described as trained from random initialization, with no specification of architecture (depth, number of filters), training iterations, loss, learning rate, or data augmentation. Since U-Net reaches a considerably higher IoU (48.82%) than scratch-trained SegFormer (37.47%), this baseline is important for interpreting whether the benefit stems from transfer learning or simply from differences in model capacity and training dynamics. To make the comparison fair and interpretable, the authors must detail the U-Net configuration and ensure that it is trained with the same data augmentation, loss function, and training schedule (where appropriate, adjusting for architecture constraints).
minor comments (6)
  1. [Section 2.4 (Training Implementation)] The learning rate schedule states a 'minimum learning rate of 1 × 10−5', which is larger than the initial learning rate of 6 × 10−6; this appears to be a typo. The intended minimum is likely 1 × 10−6, and the sentence should be corrected for clarity.
  2. [Section 2.4 (Training Implementation)] The warm-up description says the learning rate 'linearly increased from a factor of 1 × 10−6 of the initial learning rate', which gives an effective starting learning rate of 6 × 10−12, an implausibly small value. If this is indeed the implementation, it should be justified; otherwise, the factor is likely misstated and should be corrected.
  3. [References] Several reference entries are incomplete or malformed, e.g., 'Q, You, Kang S, Aguilar E, et al.' and 'F, Lutz A, Immerzeel W W, Shrestha A B, et al.' The author names and journal/volume/page fields should be corrected to the standard bibliographic format to ensure verifiability.
  4. [Section 4.2 (Limitations)] The comparison to LoveDA ('e.g., 70% IoU on LoveDA') is vague and lacks a proper citation or context. If the authors intend to benchmark against other remote sensing segmentation datasets, they should provide the source and the exact experimental conditions; otherwise, the statement should be removed.
  5. [Section 2.1 (Target Domain Dataset)] The text states that GF-2 has a resolution 'consistent with GF-7', but GF-2 typically has 1m panchromatic/4m multispectral while GF-7 has 0.8m/3.2m. If the images were resampled to a common ground sample distance, please state so explicitly; otherwise, this is misleading.
  6. [Abstract and Section 1] The phrase 'pre-train-fine-tune paradigm' is used without citing classic references for transfer learning in remote sensing (e.g., Pan and Yang 2010 or Yosinski et al. 2014 are already cited later). Consider referencing these at first use to make the connection explicit.

Circularity Check

1 steps flagged · score 6.0 of 10

The headline IoU gain (25.50% to 64.84%) is measured on a target validation set built from the same overlapping 512-by-512, stride-128 crop pool as the training set, so the claimed transfer-learning 'prediction' is largely a re-measurement of the in-sample fit rather than an out-of-sample demonstration.

  1. fitted input called prediction [Section 2.2 (Sliding Window Cropping and the 9:1 split) feeding Section 3.2, Table 2]
    "a 512×512 pixel window (Fig. 3b) is moved across the image with a stride of 128 pixels (i.e., a 25% overlap)... both datasets were randomly split into training and validation sets at a 9:1 ratio.... The fine-tuned model (A2) achieved the highest performance, with a water IoU of 64.84%."

    Stage 2 fits Model A2 to 162 target-domain training patches, and the reported 64.84% water IoU is then cited as evidence that the two-stage strategy overcomes domain shift and data scarcity. But by the paper's own construction, the 512×512 window slides with a 128-pixel stride, so adjacent crops overlap by 75% of their area, and the 'random' 9:1 split is applied to the 180-patch pool with no stated image-level or spatial separation. A validation patch therefore typically shares at least 75% of its pixels with a neighboring training patch drawn from the same GF-2 image, meaning the model was fitted on the majority of each 'held-out' patch.

full rationale

The paper's derivation chain is otherwise self-contained: Model A1 is pre-trained on the external source dataset (empirical), Model A2 is fine-tuned on target patches (empirical), and all model comparisons (direct transfer, scratch SegFormer, scratch U-Net) are measured rather than derived from fitted constants. No fitted parameter is renamed as a prediction, and no self-citations carry the argument; the citations to SegFormer, MixFormer, and transfer-learning surveys are external and non-load-bearing. The one structural circularity is the target validation protocol: because crops overlap by 75% of their area and the 9:1 split is made at the patch level with no stated image-level separation, the validation IoU reported in Section 3.2 is substantially an in-sample measurement. The paper's Section 4.2 limitation statement ('This is partly attributable to the limited size of the target domain dataset') flags small-sample issues but omits this overlap-induced leakage, which directly affects the load-bearing 25.50%-vs-64.84% comparison; the asserted 'objective and consistent evaluation' in Section 2.2 is therefore unsupported. The geospatial concentration claim (over 80% of water in under 20% of channel length) is a descriptive statistic computed from the final model's output and is not used to set any model parameter, so it is measurement-dependent rather than circular. Net: the central generalization claim partially reduces to in-sample fit by construction, warranting score 6; the residual empirical content (fine-tuning beats baselines on the observed overlapping patches) is real but does not establish domain-shift transfer as claimed.

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

No fitted parameters were used to produce the central transfer-learning comparison; training hyperparameters are standard. The main scientific claims rest on assumptions about label quality, split representativeness, and the patch-selection strategy, all documented above.

assumptions (3)
  • domain assumption Ground-truth labels for Dataset B are accurate and align with the water bodies visible in GF-2 imagery.
    All metrics in Table 2 treat the target validation labels as correct; any labeling errors directly change reported IoU. Invoked in Section 2.2 when constructing training and validation splits.
  • domain assumption The 9:1 random split of the 180 target patches produces a representative and stable validation set.
    The central performance comparison rests on this split; with 18 validation patches, the assumption is strong. Invoked in Section 2.2.
  • domain assumption Retaining only patches that contain water yields a training and validation distribution representative of the operational full-scene task.
    Water-free patches are discarded, so the reported IoU may not reflect performance on full scenes. Invoked in Section 2.2, Sliding Window Cropping.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Transfer Learning-Based Method for Water Body Segmentation in Remote Sensing Imagery: A Case Study of the Zhada Tulin Area." pith.science (2026). https://pith.science/paper/BCPYKRUZ

@misc{pith2026250710084,
  author       = {Pith},
  title        = {Pith review of: A Transfer Learning-Based Method for Water Body Segmentation in Remote Sensing Imagery: A Case Study of the Zhada Tulin Area},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BCPYKRUZ}},
  note         = {Machine review of arXiv:2507.10084}
}
read the original abstract

The Tibetan Plateau, known as the Asian Water Tower, faces significant water security challenges due to its high sensitivity to climate change. Advancing Earth observation for sustainable water monitoring is thus essential for building climate resilience in this region. This study proposes a two-stage transfer learning strategy using the SegFormer model to overcome domain shift and data scarcit--key barriers in developing robust AI for climate-sensitive applications. After pre-training on a diverse source domain, our model was fine-tuned for the arid Zhada Tulin area. Experimental results show a substantial performance boost: the Intersection over Union (IoU) for water body segmentation surged from 25.50% (direct transfer) to 64.84%. This AI-driven accuracy is crucial for disaster risk reduction, particularly in monitoring flash flood-prone systems. More importantly, the high-precision map reveals a highly concentrated spatial distribution of water, with over 80% of the water area confined to less than 20% of the river channel length. This quantitative finding provides crucial evidence for understanding hydrological processes and designing targeted water management and climate adaptation strategies. Our work thus demonstrates an effective technical solution for monitoring arid plateau regions and contributes to advancing AI-powered Earth observation for disaster preparedness in critical transboundary river headwaters.

Figures

Figures reproduced from arXiv: 2507.10084 by the authors.

Figure 1
Figure 1. Geographic location map of the study areas, highlighting the target domain in the Zhada Tulin area. the ’pre-train and fine-tune’ paradigm using ImageNet (Krizhevsky, Sutskever, and Hinton 2017) is standard, its direct application to remote sensing yields limited results due to the unique characteristics of satellite imagery (Tuia, Persello, and Bruzzone 2016). Therefore, designing transfer learning strategies tailo… view at source ↗
Figure 2
Figure 2. The overall technical framework of the two-stage transfer learning strategy. 5 [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Diagram of the data preprocessing pipeline: (a) Original remote sensing image; (b) Sliding window cropping into 512×512 patches; (c)-(f) Data augmentation on training samples. Sliding Window Cropping. To standardize the input size and effectively augment the number of samples, this study employs an overlapping sliding window method to synchronously crop the original images and their corresponding vector label maps. … view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Architecture of SegFormer, including a hierarchical Transformer encoder and an all-MLP decoder (Xie et al. 2021). The encoder of SegFormer employs the Mix Transformer (MiT) (Cui et al. 2022) series as its backbone; this study specifically uses the MiT-B5 version. Unlik…
Figure 5
Figure 5. Figure 5: Segmentation results of the foundational model A1 on a typical scene from the source dataset A: (a) Original Image, (b) Ground Truth, (c) Prediction of Model A1 [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Segmentation comparison of different models on a typical scene from the target dataset B: (a) Orig￾inal Image, (b) Ground Truth, (c) Prediction of Model A1 (Direct Transfer), (d) Prediction of Model B-scratch (SegFormer), (e) Prediction of Model A2 (Fine-tuned). The fi…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 46 canonical work pages

  1. [1]

    Badrinarayanan, Vijay, Alex Kendall, and Roberto Cipolla. 2017. ``Segnet: A deep convolutional encoder-decoder architecture for image segmentation.'' IEEE transactions on pattern analysis and machine intelligence 39 (12): 2481--2495

    work page 2017

  2. [2]

    Chen, Fahu, Chengbang An, Jiawu Zhang, et al. 2009. `` Zhongguo xibu jin qiannianlai shirundu bianhua de ruogan tezheng [Several characteristics of humidity changes in Western China over the past millennium].'' Earth Science Frontiers 16 (6): 11--19. (in Chinese)

    work page 2009

  3. [3]

    Chen, Liang-Chieh, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille. 2017. ``Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs.'' IEEE transactions on pattern analysis and machine intelligence 40 (4): 834--848

    work page 2017

  4. [4]

    Chen, Liang-Chieh, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. 2018. ``Encoder-decoder with atrous separable convolution for semantic image segmentation.'' In Proceedings of the European conference on computer vision (ECCV), 801--818

    work page 2018

  5. [5]

    Chen, Yiyu. 2015. Qingzang gaoyuan huanjing bianhua kexue pinggu: Guoqu, xianzai yu weilai [Scientific assessment of environmental change on the Qinghai-Tibet Plateau: Past, present, and future] . Science Press. (in Chinese)

    work page 2015

  6. [6]

    Cui, Yutao, Cheng Jiang, Limin Wang, and Gangshan Wu. 2022. ``Mixformer: End-to-end tracking with iterative mixed attention.'' In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 13608--13618

    work page 2022

  7. [7]

    Ding, Lin, Dalai Zhong, Yusheng Pan, et al. 2006. `` Qingzang gaoyuan de longsheng guocheng jiqi jizhi taolun [Uplift process and mechanism of the Tibetan Plateau].'' Science in China Series D: Earth Sciences 36 (9): 785--805. (in Chinese)

    work page 2006

  8. [8]

    Dosovitskiy, Alexey, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, et al. 2020. ``An image is worth 16x16 words: Transformers for image recognition at scale.'' arXiv preprint arXiv:2010.11929

    arXiv 2020

Show all 55 references
  1. [9]

    F, Lutz A, Immerzeel W W, Shrestha A B, et al. 2014. ``Consistent increase in High Asia's runoff due to increasing glacier melt and precipitation.'' Nature Climate Change 4 (7): 587--592

    2014

  2. [10]

    Fang, Xiaomin, Yongxiang Han, Yuzhen Ma, et al. 2003. `` Qingzang gaoyuan dongbeiyuan wan xinshengdai gouzao longsheng de chenji jilu [Sedimentary records of Late Cenozoic tectonic uplift in the northeastern margin of the Tibetan Plateau].'' Science in China Series D: Earth Sc...

    2003

  3. [11]

    Ganin, Yaroslav, and Victor Lempitsky. 2015. ``Unsupervised domain adaptation by backpropagation.'' In International conference on machine learning, 1180--1189. PMLR

    2015

  4. [12]

    Ghifary, Muhammad, W Bastiaan Kleijn, Mengjie Zhang, David Balduzzi, and Wen Li. 2016. ``Deep reconstruction-classification networks for unsupervised domain adaptation.'' In Computer Vision--ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11--14, 2016,...

    2016

  5. [13]

    Goodfellow, Ian, Yoshua Bengio, and Aaron Courville. 2016. Deep Learning. Cambridge, MA: MIT Press

    2016

  6. [14]

    Hoffman, Judy, Eric Tzeng, Taesung Park, Jun-Yan Zhu, Phillip Isola, Kate Saenko, Alexei Efros, and Trevor Darrell. 2018. ``Cycada: Cycle-consistent adversarial domain adaptation.'' In International conference on machine learning, 1989--1998. Pmlr

    2018

  7. [15]

    Huete, Alfredo R. 1988. ``A soil-adjusted vegetation index (SAVI).'' Remote sensing of environment 25 (3): 295--309

    1988

  8. [16]

    Krizhevsky, Alex, Ilya Sutskever, and Geoffrey E Hinton. 2017. ``ImageNet classification with deep convolutional neural networks.'' Communications of the ACM 60 (6): 84--90

    2017

  9. [17]

    LeCun, Yann, L \'e on Bottou, Yoshua Bengio, and Patrick Haffner. 2002. ``Gradient-based learning applied to document recognition.'' Proceedings of the IEEE 86 (11): 2278--2324

    2002

  10. [18]

    Li, Bingyuan, Baotian Pan, Weiming Cheng, et al. 2013. `` Zhongguo dimao quhua xinlun [A new discussion on geomorphological regionalization of China].'' Acta Geographica Sinica 68 (3): 291--306. (in Chinese)

    2013

  11. [19]

    Li, Rui, Shunyi Zheng, Ce Zhang, Chenxi Duan, Jianlin Su, Libo Wang, and Peter M Atkinson. 2021. ``Multiattention network for semantic segmentation of fine-resolution remote sensing images.'' IEEE Transactions on Geoscience and Remote Sensing 60: 1--13

    2021

  12. [20]

    Liu, Ze, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. 2021. ``Swin transformer: Hierarchical vision transformer using shifted windows.'' In Proceedings of the IEEE/CVF international conference on computer vision, 10012--10022

    2021

  13. [21]

    Long, Mingsheng, Yue Cao, Jianmin Wang, and Michael Jordan. 2015. ``Learning transferable features with deep adaptation networks.'' In International conference on machine learning, 97--105. PMLR

    2015

  14. [22]

    Loshchilov, Ilya, and Frank Hutter. 2017. ``Decoupled weight decay regularization.'' arXiv preprint arXiv:1711.05101

    2017 arXiv

  15. [23]

    Ma, Lei, Yu Liu, Xueliang Zhang, Yuanxin Ye, Gaofei Yin, and Brian Alan Johnson. 2019. ``Deep learning in remote sensing applications: A meta-analysis and review.'' ISPRS journal of photogrammetry and remote sensing 152: 166--177

    2019

  16. [24]

    McFeeters, Stuart K. 1996. ``The use of the Normalized Difference Water Index (NDWI) in the delineation of open water features.'' International journal of remote sensing 17 (7): 1425--1432

    1996

  17. [25]

    Milletari, Fausto, Nassir Navab, and Seyed-Ahmad Ahmadi. 2016. ``V-net: Fully convolutional neural networks for volumetric medical image segmentation.'' In 2016 fourth international conference on 3D vision (3DV), 565--571. Ieee

    2016

  18. [26]

    Otsu, Nobuyuki, et al. 1975. ``A threshold selection method from gray-level histograms.'' Automatica 11 (285-296): 23--27

    1975

  19. [27]

    Pan, Sinno Jialin, and Qiang Yang. 2010. ``A Survey on Transfer Learning.'' IEEE Transactions on Knowledge and Data Engineering 22 (10): 1345--1359

    2010

  20. [28]

    Paszke, A. 2019. ``Pytorch: An imperative style, high-performance deep learning library.'' arXiv preprint arXiv:1912.01703

    2019 arXiv

  21. [29]

    Pei, Zhongyi, Zhangjie Cao, Mingsheng Long, and Jianmin Wang. 2018. ``Multi-adversarial domain adaptation.'' In Proceedings of the AAAI conference on artificial intelligence, Vol. 32

    2018

  22. [30]

    Q, You, Kang S, Aguilar E, et al. 2011. ``Changes in daily climate extremes in China and their connection to the large scale atmospheric circulation during 1961–2003.'' Climate Dynamics 36 (11-12): 2399--2417

    2011

  23. [31]

    Q, Zhang, Singh V P, Li J, et al. 2023. ``Oceanic climate changes threaten the sustainability of Asia's water tower.'' Nature 615: 87--93

    2023

  24. [32]

    Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. 2015. ``U-net: Convolutional networks for biomedical image segmentation.'' In Medical image computing and computer-assisted intervention--MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, procee...

    2015

  25. [33]

    Rouse Jr, JW, RH Haas, JA Schell, and DW Deering. 1973. ``Paper a 20.'' In Third earth resources technology satellite-1 symposium: The proceedings of a symposium held by Goddard space flight center at Washington, DC on, Vol. 351, 309

    1973

  26. [34]

    Saito, Kuniaki, Kohei Watanabe, Yoshitaka Ushiku, and Tatsuya Harada. 2018. ``Maximum classifier discrepancy for unsupervised domain adaptation.'' In Proceedings of the IEEE conference on computer vision and pattern recognition, 3723--3732

    2018

  27. [35]

    Tuia, Devis, Claudio Persello, and Lorenzo Bruzzone. 2016. ``Domain adaptation for the classification of remote sensing data: An overview of recent advances.'' IEEE geoscience and remote sensing magazine 4 (2): 41--57

    2016

  28. [36]

    Tzeng, Eric, Judy Hoffman, Kate Saenko, and Trevor Darrell. 2017. ``Adversarial discriminative domain adaptation.'' In Proceedings of the IEEE conference on computer vision and pattern recognition, 7167--7176

    2017

  29. [37]

    Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, ukasz Kaiser, and Illia Polosukhin. 2017. ``Attention is all you need.'' Advances in neural information processing systems 30

    2017

  30. [38]

    W, Immerzeel W, van Beek L P H, and Bierkens M F P. 2010. ``Climate change will affect the Asian water towers.'' Science 328 (5984): 1382--1385

    2010

  31. [39]

    Wang, Junjue, Zhuo Zheng, Ailong Ma, Xiaoyan Lu, and Yanfei Zhong. 2021. ``LoveDA: A remote sensing land-cover dataset for domain adaptive semantic segmentation.'' arXiv preprint arXiv:2110.08733

    2021 arXiv

  32. [40]

    Wang, Libo, Rui Li, Ce Zhang, Shenghui Fang, Chenxi Duan, Xiaoliang Meng, and Peter M Atkinson. 2022. ``UNetFormer: A UNet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery.'' ISPRS Journal of Photogrammetry and Remote Sensing 190: 196--214

    2022

  33. [41]

    Wang, Xiaolong, Ross Girshick, Abhinav Gupta, and Kaiming He. 2018. ``Non-local neural networks.'' In Proceedings of the IEEE conference on computer vision and pattern recognition, 7794--7803

    2018

  34. [42]

    Xie, Enze, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M Alvarez, and Ping Luo. 2021. ``SegFormer: Simple and efficient design for semantic segmentation with transformers.'' Advances in neural information processing systems 34: 12077--12090

    2021

  35. [43]

    Xu, Hanqiu. 2006. ``Modification of normalised difference water index (NDWI) to enhance open water features in remotely sensed imagery.'' International journal of remote sensing 27 (14): 3025--3033

    2006

  36. [44]

    Yosinski, Jason, Jeff Clune, Yoshua Bengio, and Hod Lipson. 2014. ``How transferable are features in deep neural networks?'' Advances in neural information processing systems 27

    2014

  37. [45]

    Zhang, Cheng, Wanshou Jiang, Yuan Zhang, Wei Wang, Qing Zhao, and Chenjie Wang. 2022. ``Transformer and CNN hybrid deep neural network for semantic segmentation of very-high-resolution remote sensing imagery.'' IEEE Transactions on Geoscience and Remote Sensing 60: 1--20

    2022

  38. [46]

    Zhang, Liangpei, Lefei Zhang, and Bo Du. 2016. ``Deep learning for remote sensing data: A technical tutorial on the state of the art.'' IEEE Geoscience and remote sensing magazine 4 (2): 22--40

    2016

  39. [47]

    Zhang, Zhilu, and Mert Sabuncu. 2018. ``Generalized cross entropy loss for training deep neural networks with noisy labels.'' Advances in neural information processing systems 31

    2018

  40. [48]

    Zhu, Dagang, Dingyi Liang, Zhenhan Wu, et al. 2007. `` Xizang Zhada pendi xinshengdai diceng xulie yu guhuanjing yanbian [Cenozoic stratigraphic sequence and paleoenvironmental evolution of the Zhada Basin, Tibet].'' Geological Bulletin of China 26 (1): 1--10. (in Chinese)

    2007

  41. [49]

    Zhu, Liping, Junbo Wang, Jianmin Bian, et al. 2019. `` Qingzang gaoyuan hupo de qihou bianhua xiangying yu huanjing bianhua zhishi [Climate change response and environmental change indication of lakes on the Tibetan Plateau].'' Acta Geographica Sinica 74 (7): 1362--1379. (in Chinese)

    2019

  42. [50]

    Zhu, Xiao Xiang, Devis Tuia, Lichao Mou, Gui-Song Xia, Liangpei Zhang, Feng Xu, and Friedrich Fraundorfer. 2017. ``Deep learning in remote sensing: A comprehensive review and list of resources.'' IEEE geoscience and remote sensing magazine 5 (4): 8--36

    2017

  43. [51]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  44. [52]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  45. [53]

    subfigure

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    1997

  46. [54]

    , " * write output.state after.block = add.period

    ENTRY address archive author booktitle chapter collaboration doi edition editor eid howpublished institution journal key lastchecked month note number numpages organization pages publisher school series title translator type url urldate volume year label extra.label sort.label...

  47. [55]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.