Pith. sign in

REVIEW 4 major objections 4 minor 31 references

Pushing Trade-Off Boundaries: Compact yet Effective Remote Sensing Change Detection

T0 review · 4 major / 4 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 FlickCD claims that a change-detection network with 1.89M parameters can match or nearly match models 10–30 times larger, cutting computational and storage overhead by more than an order of magnitude.

desk verdict Genuinely useful lightweight change-detection model; efficiency gains are secure, accuracy claims need error bars and a corrected abstract. read the letter →

arxiv 2506.21109 v2 pith:GESS6P7H submitted 2025-06-26 cs.CV cs.LG

Luosheng Xu , Dalin Zhang , Zhaohui Song This is my paper · ORCID
classification cs.CVcs.LG MSC 68T0768T4568U10
keywords remotesensingchangedetectionlightweightdeeplearningdifference-awareattentionshiftedwindowself-attentionefficientglobalbi-temporalimageanalysison-satelliteprocessing
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 claims that a compact change-detection network, FlickCD, can match or nearly match state-of-the-art models that are ten to thirty times larger while cutting parameter count and computation by more than an order of magnitude. On four public remote-sensing change-detection benchmarks, the 1.89M-parameter model reports the best F1 on two (SYSU, WHU) and trails the strongest competitor, ChangeMamba, by roughly a point or less on the other two (CDD, LEVIR+). The paper ties both results to the same design choices: an Enhanced Difference Module that suppresses irrelevant differences such as lighting shifts, plus a decoder that pairs cheap local-window attention with downsampled global attention. If the measurements hold, on-satellite change detection would no longer require sacrificing accuracy for a deployable model size.

What carries the argument

The load-bearing object is the Enhanced Difference Module (EDM) and, inside it, the inverted difference-aware attention identity. The module projects both temporal feature maps into a shared space, computes the dot-product similarity $M$ between corresponding positions with $M_{ij} = q_{ij} \cdot k_{ij} / \sqrt{d_k}$, and applies the mask $M' = \text{Sigmoid}(-M)$ element-wise to the absolute-difference value $V = |f'_1 - f'_2|$, so the output is $D = M' \odot V$. Ordinary attention amplifies agreement; this inverted form amplifies disagreement, with the sigmoid bounding the mask and the $\sqrt{d_k}$ factor preventing saturation from large dot products. The second mechanism is the Local-Global Fusion Block, which combines Shifted Window Self-Attention (SWSA), attention over moving windows whose overlap is controlled by a stride, with Efficient Global Self-Attention (EGSA), where keys and values are downsampled to patch level so global context is cheap. Together these modules let a three-stage RepViT encoder feed a decoder that the paper claims preserves both coarse and fine-grained change boundaries at a fraction of the cost of heavier models.

What would settle it

Re-run the four benchmark comparisons with several random seeds and report the spread of F1 scores with per-dataset hyperparameters fixed. If the deficits on CDD and LEVIR+ fall inside the seed-to-seed variance, for example if FlickCD and ChangeMamba produce overlapping ranges, then the 'minor (<1% F1) trade-off' claim is not established; if the gaps persist across seeds, the trade-off claim holds.

Watch

Extended reading notes

Core claim

The paper's central claim, stated on its own terms, is that the accuracy-efficiency trade-off in change detection is not fixed: a deliberately small model can sit at the frontier if the difference signal is computed well. FlickCD rests on an inversion of the standard attention idea. Where ordinary attention boosts tokens whose vectors agree, the Enhanced Difference Module computes a similarity score between the two temporal feature maps, inverts it, and uses the result to weight the absolute difference between the maps, so pixels whose bi-temporal features genuinely diverge are amplified while lighting and weather noise are pressed down. The decoder then refines this weighted difference with two cheap attention mechanisms: Shifted Window Self-Attention (SWSA), which confines attention to local windows with an adjustable stride so windows can overlap, and Efficient Global Self-Attention (EGSA), which downsamples keys and values to one token per window so each pixel can consult the whole image at low cost. The paper further argues that the encoder's fourth stage adds little and can be dropped, and reports that the resulting 1.89M-parameter, 4.16-GFLOP model achieves the best F1 on SYSU and WHU and comes within roughly a point of the leading model on CDD and LEVIR+.

Load-bearing premise

The claim that the accuracy loss is only 'minor' assumes that the reported F1 gaps, most under one percentage point against ChangeMamba, reflect real differences rather than the noise of a single training run with window sizes, strides, and learning rates tuned per dataset.

Editorial extensions

If this is right

  • On SYSU and WHU, FlickCD reports the highest F1 while using about 3.5% of ChangeMamba's parameters, which implies that model size is not what sets the accuracy ceiling on these benchmarks.
  • On CDD and LEVIR+, FlickCD lands within roughly a point of ChangeMamba while running at less than a third of its per-sample inference time, so a near-frontier detector can fit in a small memory budget.
  • The ablations show EDM raises F1 on every dataset, and removing either SWSA or EGSA costs accuracy on the many-region building datasets WHU and LEVIR+ while combining both helps everywhere, implying local and global context are complementary rather than interchangeable.
  • Dropping the fourth encoder stage and using depthwise-separable projections each halve the parameter count with no measured F1 loss, implying that a large share of the parameters in bigger models may be doing non-essential work for this task.

Reading between the lines

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

  • The paper's dataset-characteristic analysis suggests a cheap, transferable recipe: complex sparse changes want larger overlapping attention windows, while regular many-region changes (building datasets) want smaller ones; other lightweight detectors could adopt this per-dataset knob without changing architecture, though the paper does not test that transfer.
  • The inverted-difference attention at the heart of EDM is a generic operator on two feature maps, so it could plausibly transfer to other bi-temporal problems such as video frame differencing or anomaly surveillance, where nuisance variation must be suppressed; this is an extension the paper leaves untested.
  • The trade-off claim rests on single-run F1 scores, so the natural next check is seed-averaged comparisons against equally tuned lightweight baselines; if those close the gap, the frontier is defined by tuning practice as much as by architecture.
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 / 4 minor

Summary. The paper proposes FlickCD, a lightweight Siamese change-detection model built on a RepViT encoder, an Enhanced Difference Module (EDM) for filtering irrelevant differences, and a decoder containing Local-Global Fusion Blocks with Sliding-Window Self-Attention (SWSA) and Efficient Global Self-Attention (EGSA). The authors report results on four change-detection benchmarks (SYSU, WHU, CDD, LEVIR+) and claim that FlickCD reduces parameter count and FLOPs by more than an order of magnitude relative to large models while achieving state-of-the-art performance or incurring only a minor (<1% F1) trade-off. The implementation code is publicly available.

Significance. If the accuracy comparisons are statistically robust, the contribution is significant: FlickCD uses 1.89M parameters and 4.16G FLOPs versus 53.99M parameters and 61.69G FLOPs for ChangeMamba, with roughly three times the throughput and one-third of the inference time, across four standard benchmarks. The module-level ablations and the dataset-characteristic analysis in Figure 4 provide useful diagnostics for why local and global attention both matter. The central weakness is that the headline accuracy claims rest on single-run F1 differences of a few tenths of a percentage point, measured under per-dataset hyperparameter tuning, so the accuracy side of the claimed trade-off is not yet established.

major comments (4)
  1. [Abstract / Section 4.2, Tables 1 and 2] The abstract's claim of a '<1% F1 accuracy trade-off' is not literally satisfied by the LEVIR+ row, where FlickCD trails ChangeMamba by 1.04 percentage points (87.34 vs 86.30); if the claim is intended as a four-dataset average, the text must say so explicitly. More importantly, the comparisons are single-run numbers with no error bars, so the 0.14 pp lead over BiFA on SYSU and the 0.25 pp lead over ChangeMamba on WHU cannot be distinguished from run-to-run noise. The authors should report multi-seed means with standard deviations or confidence intervals for at least the key comparisons, or temper the SOTA and '<1%' claims accordingly.
  2. [Section 4.1, Experiment Settings] The SWSA window sizes and strides, learning rates, and training epochs are tuned individually for each dataset, and Table 4 shows that module configuration choices can shift F1 by more than 1 pp (e.g., SYSU from 80.41 to 83.97). Because the central claim depends on sub-1-point margins over the strongest baselines, the per-dataset tuning protocol must be specified (for example, selection on a validation split) and ideally the hyperparameters would be held fixed across datasets or shown to be insensitive. Without this, the comparison is biased in favor of the proposed model and the measured lead is not a fair test of the trade-off claim.
  3. [Section 4.3, Table 5] The text states that the two lightweight strategies (removing depthwise convolutions and using only three encoder stages) produce 'no performance degradation observed across the four datasets,' but Table 5 shows the opposite: both variants are lower than the base model on every dataset, for example SYSU 83.97 versus 83.51 and 82.67, and LEVIR+ 86.30 versus 85.68 and 85.99. The sentence should be corrected, and the magnitude of the degradation -- even if small -- should be interpreted in light of the absence of error bars. This matters because the conclusion that the information loss is 'non-critical' is directly based on these numbers.
  4. [Section 4.2, Quantitative Results] The statement that FlickCD 'achieves the highest performance in SYSU and WHU datasets' is not supported by statistically meaningful margins: the lead over the second-best model is 0.14 pp on SYSU and 0.25 pp on WHU, both from single runs. The authors should either provide paired evaluations across multiple seeds, which would allow a reader to judge whether these margins are reproducible, or explicitly present FlickCD as tied with the top models on these datasets rather than as state-of-the-art.
minor comments (4)
  1. [Throughout] The model name is inconsistently typeset as 'FlickCD' and 'FLICKCD'; please standardize the notation.
  2. [Section 3.2, Eq. (1)] The notation 'M ⊃ m_ij' is nonstandard and difficult to parse; use 'M = [m_ij]' or 'with entries m_ij' and ensure the spatial indices i and j are consistent with the stated feature-map resolution W/2^{i+1} × H/2^{i+1}.
  3. [Table 3] The throughput and inference-time measurements lack hardware, software framework, input resolution, and batch-size details; without these, the numbers are not reproducible and should be supplemented.
  4. [Figure 1 caption] The description of relative performance normalization is ambiguous; please clarify whether the F1 normalization is performed separately per dataset or globally across all datasets.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: FlickCD's claims are empirical measurements on external benchmarks with no self-citation or fitted-input-as-prediction structure.

full rationale

FlickCD is an empirical architecture paper; there is no derivation chain whose output is pre-supplied by its inputs. The reported results are F1/IoU and resource measurements on four external benchmarks (SYSU, WHU, LEVIR+, CDD), with comparisons against previously published models. The per-dataset SWSA window/stride and learning-rate settings in Section 4.1 are hyperparameter choices, not fitted parameters that determine the headline quantities; the central claims are the measured accuracy and the parameter/FLOP counts, which do not reduce to any fitted value. The paper contains no self-citations to the authors' prior work, and the components (RepViT encoder, EDM, SWSA, EGSA) are described by explicit equations and tested via ablations against the same benchmarks. The abstract's '<1% F1 trade-off' is numerically inconsistent with the LEVIR+ row if read literally, but that is a reporting/statistical robustness issue, not a circularity of the kind defined here.

Assumptions & free parameters 6 free parameters · 3 assumptions · 0 invented entities

The listed hyperparameters are chosen per dataset and affect the reported F1 scores. The pretrained encoder and benchmark protocols are taken as given. No new physical entities are introduced.

free parameters (6)
  • SWSA window sizes (SYSU, CDD) = 8, 8, 16
    Set larger for datasets with high shape complexity; stated in Experiment Settings, Section 4.1.
  • SWSA strides (SYSU, CDD) = 4, 4, 8
    Paired with the window sizes to allow overlapping receptive fields; Section 4.1.
  • SWSA window sizes and strides (WHU) = 4, 4, 8 (window and stride)
    Chosen for regular building changes; Section 4.1.
  • SWSA window sizes and strides (LEVIR+) = 4, 8, 8 (window and stride)
    Chosen for regular building changes; Section 4.1.
  • Learning rate per dataset = 0.0005 (SYSU, CDD, LEVIR+); 0.0002 (WHU)
    Adjusted per dataset; Section 4.1.
  • Training epochs per dataset = 100 (SYSU, WHU); 200 (LEVIR+); 250 (CDD)
    Adjusted per dataset; Section 4.1.
assumptions (3)
  • domain assumption ImageNet-pretrained RepViT features transfer to remote sensing change detection.
    The encoder is initialized with pretrained weights (Section 3.1); the central results depend on this transfer being effective.
  • domain assumption The four benchmarks (SYSU, WHU, CDD, LEVIR+) are representative of change detection tasks.
    Generalization claims are made from these datasets without cross-dataset validation; Section 4.1.
  • domain assumption The reported metrics are computed with standard evaluation protocols.
    No code or protocol details are given to verify the exact evaluation; Section 4.2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pushing Trade-Off Boundaries: Compact yet Effective Remote Sensing Change Detection." pith.science (2026). https://pith.science/paper/GESS6P7H

@misc{pith2026250621109,
  author       = {Pith},
  title        = {Pith review of: Pushing Trade-Off Boundaries: Compact yet Effective Remote Sensing Change Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GESS6P7H}},
  note         = {Machine review of arXiv:2506.21109}
}
read the original abstract

Remote sensing change detection is essential for monitoring urban expansion, disaster assessment, and resource management, offering timely, accurate, and large-scale insights into dynamic landscape transformations. While deep learning has revolutionized change detection, the increasing complexity and computational demands of modern models have not necessarily translated into significant accuracy gains. Instead of following this trend, this study explores a more efficient approach, focusing on lightweight models that maintain high accuracy while minimizing resource consumption, which is an essential requirement for on-satellite processing. To this end, we propose FlickCD, which means quick flick then get great results, pushing the boundaries of the performance-resource trade-off. FlickCD introduces an Enhanced Difference Module (EDM) to amplify critical feature differences between temporal phases while suppressing irrelevant variations such as lighting and weather changes, thereby reducing computational costs in the subsequent change decoder. Additionally, the FlickCD decoder incorporates Local-Global Fusion Blocks, leveraging Shifted Window Self-Attention (SWSA) and Efficient Global Self-Attention (EGSA) to effectively capture semantic information at multiple scales, preserving both coarse- and fine-grained changes. Extensive experiments on four benchmark datasets demonstrate that FlickCD reduces computational and storage overheads by more than an order of magnitude while achieving state-of-the-art (SOTA) performance or incurring only a minor (<1% F1) accuracy trade-off. The implementation code is publicly available at https://github.com/xulsh8/FlickCD.

Figures

Figures reproduced from arXiv: 2506.21109 by the authors.

Figure 1
Figure 1. Comparison of parameters and relative performance between FlickCD (Ours) and others. The relative [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. (a) Overall architecture of the FlickCD model. (b) Structural illustration of the Enhanced Difference Module [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. (a)Structural diagram of the Sliding Window Self-Attention (SWSA) module. (b)Detailed architecture of the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Changed region characteristics in four datasets. For each dataset, samples are divided by the number of [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Visualization results on the WHU and LEVIR+ dataset, where white represents true positive, black represents [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Visualization results on the SYSU dataset, showcasing the outputs of other models, FlickCD with EDM, and [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Visualization of intermediate feature maps and attention heatmaps: (a) Difference map before EDM; (b) [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 19 canonical work pages

  1. [1]

    Dual-tasks siamese transformer framework for building damage assessment

    Hongruixuan Chen, Edoardo Nemni, Sofia Vallecorsa, Xi Li, Chen Wu, and Lars Bromley. Dual-tasks siamese transformer framework for building damage assessment. InIGARSS 2022-2022 IEEE International Geoscience and Remote Sensing Symposium, pages 1600–1603. IEEE, 2022

    work page 2022

  2. [2]

    Zhuo Zheng, Yanfei Zhong, Junjue Wang, Ailong Ma, and Liangpei Zhang. Building damage assessment for rapid disaster response with a deep object-based semantic change detection framework: From natural disasters to man-made disasters.Remote Sensing of Environment, 265:112636, 2021

    work page 2021

  3. [3]

    Haonan Guo, Qian Shi, Andrea Marinoni, Bo Du, and Liangpei Zhang. Deep building footprint update network: A semi-supervised method for updating existing building footprint from bi-temporal remote sensing images.Remote Sensing of Environment, 264:112589, 2021

    work page 2021

  4. [4]

    Kangkai Lou, Mengmeng Li, Fashuai Li, and Xiangtao Zheng. Integrating local–global structural interaction using siamese graph neural network for urban land use change detection from vhr satellite images.IEEE Transactions on Geoscience and Remote Sensing, 2024

    work page 2024

  5. [5]

    Agricultural remote sensing big data: Management and applications.Journal of Integrative Agriculture, 17(9):1915–1931, 2018

    Yanbo Huang, Zhong-xin Chen, YU Tao, Xiang-zhi Huang, and Xing-fa Gu. Agricultural remote sensing big data: Management and applications.Journal of Integrative Agriculture, 17(9):1915–1931, 2018

    work page 1915

  6. [6]

    A review of remote sensing for environmental monitoring in china.Remote Sensing, 12(7):1130, 2020

    Jun Li, Yanqiu Pei, Shaohua Zhao, Rulin Xiao, Xiao Sang, and Chengye Zhang. A review of remote sensing for environmental monitoring in china.Remote Sensing, 12(7):1130, 2020

    work page 2020

  7. [7]

    Relating CNN-Transformer Fusion Network for Change Detection

    Yuhao Gao, Gensheng Pei, Mengmeng Sheng, Zeren Sun, Tao Chen, and Yazhou Yao. Relating cnn-transformer fusion network for change detection.arXiv preprint arXiv:2407.03178, 2024

    work page Pith review arXiv 2024

  8. [8]

    Changemamba: Remote sensing change detection with spatio-temporal state space model.IEEE Transactions on Geoscience and Remote Sensing, 2024

    Hongruixuan Chen, Jian Song, Chengxi Han, Junshi Xia, and Naoto Yokoya. Changemamba: Remote sensing change detection with spatio-temporal state space model.IEEE Transactions on Geoscience and Remote Sensing, 2024

    work page 2024

Show all 31 references
  1. [9]

    Ultra- lightweight spatial–spectral feature cooperation network for change detection in remote sensing images.IEEE Transactions on Geoscience and Remote Sensing, 61:1–14, 2023

    Tao Lei, Xinzhe Geng, Hailong Ning, Zhiyong Lv, Maoguo Gong, Yaochu Jin, and Asoke K Nandi. Ultra- lightweight spatial–spectral feature cooperation network for change detection in remote sensing images.IEEE Transactions on Geoscience and Remote Sensing, 61:1–14, 2023

    2023

  2. [10]

    Tinycd: A (not so) deep learning model for change detection.Neural Computing and Applications, 35(11):8471–8486, 2023

    Andrea Codegoni, Gabriele Lombardi, and Alessandro Ferrari. Tinycd: A (not so) deep learning model for change detection.Neural Computing and Applications, 35(11):8471–8486, 2023

    2023

  3. [11]

    Remote sensing image change detection with transformers.IEEE Transactions on Geoscience and Remote Sensing, 60:1–14, 2021

    Hao Chen, Zipeng Qi, and Zhenwei Shi. Remote sensing image change detection with transformers.IEEE Transactions on Geoscience and Remote Sensing, 60:1–14, 2021

    2021

  4. [12]

    A transformer-based siamese network for change detection

    Wele Gedara Chaminda Bandara and Vishal M Patel. A transformer-based siamese network for change detection. InIGARSS 2022-2022 IEEE International Geoscience and Remote Sensing Symposium, pages 207–210. IEEE, 2022

    2022

  5. [13]

    Multi-scale change-aware transformer for remote sensing image change detection

    Huan Chen, Tingfa Xu, Zhenxiang Chen, Peifu Liu, Huiyan Bai, and Jianan Li. Multi-scale change-aware transformer for remote sensing image change detection. InProceedings of the 32nd ACM International Conference on Multimedia, pages 2992–3000, 2024. 11

    2024

  6. [14]

    Repvit: Revisiting mobile cnn from vit perspective

    Ao Wang, Hui Chen, Zijia Lin, Jungong Han, and Guiguang Ding. Repvit: Revisiting mobile cnn from vit perspective. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15909–15920, 2024

    2024

  7. [15]

    Fully convolutional siamese networks for change detection

    Rodrigo Caye Daudt, Bertr Le Saux, and Alexandre Boulch. Fully convolutional siamese networks for change detection. In2018 25th IEEE international conference on image processing (ICIP), pages 4063–4067. IEEE, 2018

    2018

  8. [16]

    Snunet-cd: A densely connected siamese network for change detection of vhr images.IEEE Geoscience and Remote Sensing Letters, 19:1–5, 2021

    Sheng Fang, Kaiyu Li, Jinyuan Shao, and Zhe Li. Snunet-cd: A densely connected siamese network for change detection of vhr images.IEEE Geoscience and Remote Sensing Letters, 19:1–5, 2021

    2021

  9. [17]

    Chenxiao Zhang, Peng Yue, Deodato Tapete, Liangcun Jiang, Boyi Shangguan, Li Huang, and Guangchao Liu. A deeply supervised image fusion network for change detection in high resolution bi-temporal remote sensing images.ISPRS Journal of Photogrammetry and Remote Sensing, 166:183...

    2020

  10. [18]

    Optical remote sensing image change detection based on attention mechanism and image difference.IEEE Transactions on Geoscience and Remote Sensing, 59(9):7296– 7307, 2020

    Xueli Peng, Ruofei Zhong, Zhen Li, and Qingyang Li. Optical remote sensing image change detection based on attention mechanism and image difference.IEEE Transactions on Geoscience and Remote Sensing, 59(9):7296– 7307, 2020

    2020

  11. [19]

    Land- cover change detection using paired openstreetmap data and optical high-resolution imagery via object-guided transformer.arXiv preprint arXiv:2310.02674, 2023

    Hongruixuan Chen, Cuiling Lan, Jian Song, Clifford Broni-Bediako, Junshi Xia, and Naoto Yokoya. Land- cover change detection using paired openstreetmap data and optical high-resolution imagery via object-guided transformer.arXiv preprint arXiv:2310.02674, 2023

    arXiv

  12. [20]

    Mobilenets: Efficient convolutional neural networks for mobile vision applications.arXiv preprint arXiv:1704.04861, 2017

    Andrew G Howard. Mobilenets: Efficient convolutional neural networks for mobile vision applications.arXiv preprint arXiv:1704.04861, 2017

    2017 arXiv

  13. [21]

    Shufflenet: An extremely efficient convolutional neural network for mobile devices

    Xiangyu Zhang, Xinyu Zhou, Mengxiao Lin, and Jian Sun. Shufflenet: An extremely efficient convolutional neural network for mobile devices. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 6848–6856, 2018

    2018

  14. [22]

    Run, don’t walk: chasing higher flops for faster neural networks

    Jierun Chen, Shiu-hong Kao, Hao He, Weipeng Zhuo, Song Wen, Chul-Ho Lee, and S-H Gary Chan. Run, don’t walk: chasing higher flops for faster neural networks. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12021–12031, 2023

    2023

  15. [23]

    Swin transformer: Hierarchical vision transformer using shifted windows

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

    2021

  16. [24]

    Axial-deeplab: Stand-alone axial-attention for panoptic segmentation

    Huiyu Wang, Yukun Zhu, Bradley Green, Hartwig Adam, Alan Yuille, and Liang-Chieh Chen. Axial-deeplab: Stand-alone axial-attention for panoptic segmentation. InEuropean conference on computer vision, pages 108–126. Springer, 2020

    2020

  17. [25]

    Mobilevit: light-weight, general-purpose, and mobile-friendly vision transformer.arXiv preprint arXiv:2110.02178, 2021

    Sachin Mehta and Mohammad Rastegari. Mobilevit: light-weight, general-purpose, and mobile-friendly vision transformer.arXiv preprint arXiv:2110.02178, 2021

    2021 arXiv

  18. [26]

    Efficientnet: Rethinking model scaling for convolutional neural networks

    Mingxing Tan and Quoc Le. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pages 6105–6114. PMLR, 2019

    2019

  19. [27]

    Qian Shi, Mengxi Liu, Shengchen Li, Xiaoping Liu, Fei Wang, and Liangpei Zhang. A deeply supervised attention metric-based network and an open aerial image dataset for remote sensing change detection.IEEE transactions on geoscience and remote sensing, 60:1–16, 2021

    2021

  20. [28]

    Shunping Ji, Shiqing Wei, and Meng Lu. Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set.IEEE Transactions on geoscience and remote sensing, 57(1):574–586, 2018

    2018

  21. [29]

    A spatial-temporal attention-based method and a new dataset for remote sensing image change detection.Remote sensing, 12(10):1662, 2020

    Hao Chen and Zhenwei Shi. A spatial-temporal attention-based method and a new dataset for remote sensing image change detection.Remote sensing, 12(10):1662, 2020

    2020

  22. [30]

    MA Lebedev, Yu V Vizilter, OV Vygolov, Vladimir A Knyaz, and A Yu Rubis. Change detection in remote sensing images using conditional adversarial networks.The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 42:565–571, 2018

    2018

  23. [31]

    Bifa: Remote sensing image change detection with bitemporal feature alignment.IEEE Transactions on Geoscience and Remote Sensing, 2024

    Haotian Zhang, Hao Chen, Chenyao Zhou, Keyan Chen, Chenyang Liu, Zhengxia Zou, and Zhenwei Shi. Bifa: Remote sensing image change detection with bitemporal feature alignment.IEEE Transactions on Geoscience and Remote Sensing, 2024. 12

    2024

Pith tools

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