pith. machine review for the scientific record. sign in

arxiv: 2604.11231 · v1 · submitted 2026-04-13 · 💻 cs.CV

Recognition: unknown

Seg2Change: Adapting Open-Vocabulary Semantic Segmentation Model for Remote Sensing Change Detection

You Su , Yonghong Song , Jingqi Chen , Zehan Wen
Authors on Pith no claims yet

Pith reviewed 2026-05-10 15:21 UTC · model grok-4.3

classification 💻 cs.CV
keywords open-vocabulary change detectionremote sensingsemantic segmentation adaptercategory-agnostic change headCA-CDD datasetland cover change
0
0 comments X

The pith

A simple adapter turns open-vocabulary semantic segmentation models into open-vocabulary change detectors for remote sensing imagery.

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

The paper shows how to adapt existing open-vocabulary semantic segmentation models to the task of detecting changes across any category in remote sensing images. It does so by building a category-agnostic change detection dataset and attaching a lightweight change head that identifies transitions without being limited to fixed classes. A reader should care because conventional change detection systems can only spot changes in the exact classes they were trained on, which limits their usefulness when new land-cover types appear. If the approach works, it means vision-language models already trained on broad vocabularies can be reused for flexible monitoring of human and ecological impacts without starting from scratch.

Core claim

We construct the category-agnostic change detection dataset CA-CDD and introduce a category-agnostic change head that detects arbitrary category transitions and indexes them to specific classes. Based on these elements we present Seg2Change, an adapter that converts pre-trained open-vocabulary semantic segmentation models into open-vocabulary change detection models. The resulting framework reaches state-of-the-art results on OVCD benchmarks.

What carries the argument

The category-agnostic change head, which attaches to a pre-trained open-vocabulary segmentation backbone, detects transitions between arbitrary categories and maps them to language-specified classes.

If this is right

  • Open-vocabulary change detection becomes feasible without retraining entire models or restricting outputs to a fixed label set.
  • Changes involving categories absent from the original training data can still be detected and named using language descriptions.
  • The adapter approach yields measurable gains on established benchmarks such as WHU-CD and SECOND.
  • Existing open-vocabulary segmentation models can be reused for change detection with only the addition of the new head and dataset.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar lightweight adapters could allow the same segmentation models to handle other remote sensing tasks such as object counting or anomaly localization.
  • The method opens the possibility of querying change maps with natural language descriptions rather than predefined class lists.
  • If domain shift proves larger than expected, targeted fine-tuning of only the change head might still preserve most of the open-vocabulary capability.
  • The framework could reduce reliance on large-scale, fully labeled change detection datasets by leveraging segmentation models already trained on diverse imagery.

Load-bearing premise

A category-agnostic change head attached to pre-trained open-vocabulary segmentation models will generalize to real-world remote sensing change detection without major domain shift or loss of open-vocabulary ability.

What would settle it

Training Seg2Change on a new remote sensing dataset with many unseen categories and observing that it fails to outperform closed-set change detectors on those categories would falsify the central claim.

Figures

Figures reproduced from arXiv: 2604.11231 by Jingqi Chen, Yonghong Song, You Su, Zehan Wen.

Figure 1
Figure 1. Figure 1: Previous paradigms vs. our paradigm. Previous paradigms rely on change proposals and are constrained by the [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Visual comparison between predefined change cate [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Annotation process of the category-agnostic change [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Category-Agnostic Change Head (CACH). Bi-temporal image features are extracted and processed through feature [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Bi-temporal Difference Fusion Module (BDFM) and Effective Difference Query Attention (EDQA) Module. [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Open-vocabulary building and land-cover change detection examples. In each group: images at T1 ( [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Open-vocabulary semantic change detection examples. In each group: images at T1 ( [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Visualization of feature outputs across our modules. [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Overall of Seg2Change. The bi-temporal remote sensing images are first fed into the category-agnostic change head [PITH_FULL_IMAGE:figures/full_fig_p013_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Example images from binary change detection datasets (WHU-CD, LEVIR-CD, DSIFN, and CLCD) and semantic [PITH_FULL_IMAGE:figures/full_fig_p015_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Visual comparison between CA-CDD category-agnostic change labels and the original coarse labels. We have improved [PITH_FULL_IMAGE:figures/full_fig_p016_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Qualitative comparison of several representative UCD and OVCD methods on three binary change detection datasets, [PITH_FULL_IMAGE:figures/full_fig_p017_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Qualitative comparison of several representative UCD and OVCD methods on CLCD. [PITH_FULL_IMAGE:figures/full_fig_p018_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Qualitative comparison of the state-of-the-art OVCD method DynamicEarth on a semantic change detection dataset, [PITH_FULL_IMAGE:figures/full_fig_p018_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Qualitative comparison of the state-of-the-art OVCD method DynamicEarth on a semantic change detection dataset, [PITH_FULL_IMAGE:figures/full_fig_p019_15.png] view at source ↗
read the original abstract

Change detection is a fundamental task in remote sensing, aiming to quantify the impacts of human activities and ecological dynamics on land-cover changes. Existing change detection methods are limited to predefined classes in training datasets, which constrains their scalability in real-world scenarios. In recent years, numerous advanced open-vocabulary semantic segmentation models have emerged for remote sensing imagery. However, there is still a lack of an effective framework for directly applying these models to open-vocabulary change detection (OVCD), a novel task that integrates vision and language to detect changes across arbitrary categories. To address these challenges, we first construct a category-agnostic change detection dataset, termed CA-CDD. Further, we design a category-agnostic change head to detect the transitions of arbitrary categories and index them to specific classes. Based on them, we propose Seg2Change, an adapter designed to adapt open-vocabulary semantic segmentation models to change detection task. Without bells and whistles, this simple yet effective framework achieves state-of-the-art OVCD performance (+9.52 IoU on WHU-CD and +5.50 mIoU on SECOND). Our code is released at https://github.com/yogurts-sy/Seg2Change.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper introduces Seg2Change, an adapter framework that repurposes pre-trained open-vocabulary semantic segmentation (OVSS) models for open-vocabulary change detection (OVCD) in remote sensing. It constructs a new category-agnostic change detection dataset (CA-CDD) and attaches a category-agnostic change head to detect and index arbitrary-class transitions. The central empirical claim is that this simple design yields SOTA OVCD results (+9.52 IoU on WHU-CD, +5.50 mIoU on SECOND) without bells and whistles, with code released.

Significance. If the reported gains are robust, the work supplies a practical route to scalable, class-agnostic change detection that overcomes the closed-set limitation of prior remote-sensing CD methods. The new CA-CDD dataset and public code release are concrete assets that lower the barrier for follow-on research. The significance is reduced, however, by the untested assumption that a lightweight change head can fully compensate for domain shift between natural-image OVSS pre-training and aerial/satellite imagery while preserving open-vocabulary indexing.

major comments (2)
  1. [Experiments] Experiments section (results on WHU-CD and SECOND): the reported +9.52 IoU and +5.50 mIoU gains are presented without error bars, statistical significance tests, or explicit confirmation that data splits and category sampling were fixed before seeing results. This directly affects the strength of the SOTA claim and the assertion that the gains arise from the proposed adapter rather than dataset-specific tuning.
  2. [Method] Method section, category-agnostic change head design: the central claim that the head enables reliable open-vocabulary indexing on remote-sensing data rests on the untested assumption that it can compensate for spectral/textural/scale domain shift from natural-image pre-training. No ablation or diagnostic is shown that measures preservation of class-agnostic feature alignment (e.g., zero-shot transfer accuracy on held-out remote-sensing categories before and after head attachment).
minor comments (2)
  1. [Abstract] Abstract: the phrase 'without bells and whistles' is repeated but the precise training protocol for the change head (frozen vs. lightly tuned backbone, loss formulation, number of epochs) is not summarized, making it hard for readers to reproduce the claimed simplicity.
  2. [Introduction] The paper should add a short related-work paragraph contrasting the new CA-CDD construction with existing change-detection benchmarks to clarify its novelty.

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for the detailed review and valuable feedback on our manuscript. We appreciate the recognition of the significance of Seg2Change and the CA-CDD dataset. We address each of the major comments below and outline the revisions we will make to strengthen the paper.

read point-by-point responses

  1. Referee: [Experiments] Experiments section (results on WHU-CD and SECOND): the reported +9.52 IoU and +5.50 mIoU gains are presented without error bars, statistical significance tests, or explicit confirmation that data splits and category sampling were fixed before seeing results. This directly affects the strength of the SOTA claim and the assertion that the gains arise from the proposed adapter rather than dataset-specific tuning.

    Authors: We agree that the absence of error bars, standard deviations, and statistical significance tests limits the robustness of the SOTA claims. In the revised manuscript we will report results averaged over at least three independent runs with different random seeds, including mean and standard deviation for IoU on WHU-CD and mIoU on SECOND. We will also add paired statistical tests (e.g., t-test) against the strongest baselines. We confirm that the official data splits of WHU-CD and SECOND were used without modification and that the category definitions and sampling strategy for CA-CDD were fixed before any training or evaluation; we will state this explicitly in the experiments section. revision: yes

  2. Referee: [Method] Method section, category-agnostic change head design: the central claim that the head enables reliable open-vocabulary indexing on remote-sensing data rests on the untested assumption that it can compensate for spectral/textural/scale domain shift from natural-image pre-training. No ablation or diagnostic is shown that measures preservation of class-agnostic feature alignment (e.g., zero-shot transfer accuracy on held-out remote-sensing categories before and after head attachment).

    Authors: We acknowledge that the manuscript does not contain a direct diagnostic measuring whether the change head preserves the open-vocabulary alignment of the pre-trained OVSS features under remote-sensing domain shift. We will add an ablation that evaluates zero-shot semantic segmentation accuracy on held-out remote-sensing categories (drawn from a withheld portion of CA-CDD) both before and after attaching the category-agnostic change head. This will quantify any degradation in class-agnostic feature alignment and support the claim that the head enables change detection without destroying the inherited open-vocabulary indexing capability. revision: yes

Circularity Check

0 steps flagged

No significant circularity; new dataset and adapter provide independent empirical content

full rationale

The paper's central claims rest on constructing a new category-agnostic dataset (CA-CDD) and designing a new change head/adapter (Seg2Change) that is then evaluated on standard external benchmarks (WHU-CD, SECOND). No equations or steps reduce by construction to fitted parameters on the test data, self-cited uniqueness theorems, or renamed prior results. The derivation chain is self-contained via novel components and reported performance gains.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The approach relies on the existence and quality of pre-trained open-vocabulary segmentation models plus the assumption that a lightweight change head can be trained on the new dataset without destroying open-vocabulary properties.

axioms (1)
  • domain assumption Pre-trained open-vocabulary segmentation models produce reliable per-pixel embeddings that can be compared across time for change detection.
    Invoked when the adapter is attached to existing models without retraining the backbone.
invented entities (1)
  • Category-agnostic change head no independent evidence
    purpose: Detects transitions between arbitrary categories and indexes them to specific classes.
    New module introduced to bridge segmentation models to change detection.

pith-pipeline@v0.9.0 · 5520 in / 1306 out tokens · 42318 ms · 2026-05-10T15:21:15.034560+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

81 extracted references · 43 canonical work pages · 6 internal anchors

  1. [1]

    Jean-Baptiste Alayrac, Jeff Donahue, Pauline Luc, Antoine Miech, Iain Barr, Yana Hasson, Karel Lenc, Arthur Mensch, Katherine Millican, Malcolm Reynolds, Roman Ring, Eliza Rutherford, Serkan Cabi, Tengda Han, Zhitao Gong, Sina Samangooei, Marianne Monteiro, Jacob L Menick, Sebastian Borgeaud, Andy Brock, Aida Nematzadeh, Sahand Sharifzadeh, Mikoł aj Bińko...

    2022

  2. [2]

    Wele Gedara Chaminda Bandara and Vishal M. Patel. 2022. A Transformer-Based Siamese Network for Change Detection. InIGARSS 2022 - 2022 IEEE International Geoscience and Remote Sensing Symposium. 207–210. doi:10.1109/IGARSS46834. 2022.9883686

    work page doi:10.1109/igarss46834 2022

  3. [3]

    Luca Bergamasco, Sudipan Saha, Francesca Bovolo, and Lorenzo Bruzzone. 2022. Unsupervised change detection using convolutional-autoencoder multiresolution features.IEEE Transactions on Geoscience and Remote Sensing60 (2022), 1–19

    2022

  4. [4]

    Francesca Bovolo and Lorenzo Bruzzone. 2007. A Theoretical Framework for Unsupervised Change Detection Based on Change Vector Analysis in the Polar Domain.IEEE Transactions on Geoscience and Remote Sensing45, 1 (2007), 218–236. doi:10.1109/TGRS.2006.885408

    work page doi:10.1109/tgrs.2006.885408 2007

  5. [5]

    Francesca Bovolo, Lorenzo Bruzzone, and Mattia Marconcini. 2008. A Novel Approach to Unsupervised Change Detection Based on a Semisupervised SVM and a Similarity Measure.IEEE Transactions on Geoscience and Remote Sensing 46, 7 (2008), 2070–2082. doi:10.1109/TGRS.2008.916643

    work page doi:10.1109/tgrs.2008.916643 2008

  6. [6]

    Lorenzo Bruzzone and Diego F Prieto. 2002. Automatic analysis of the difference image for unsupervised change detection.IEEE Transactions on Geoscience and Remote sensing38, 3 (2002), 1171–1182

    2002

  7. [7]

    F. P. Kressler C. A. Mucher, K. T. Steinnocher and C. Heunks. 2000. Land cover characterization and change detection for environmental monitoring of pan- Europe.International Journal of Remote Sensing21, 6-7 (2000), 1159–1181. doi:10. 1080/014311600210128

    2000

  8. [8]

    Nicolas Carion, Laura Gustafson, Yuan-Ting Hu, Shoubhik Debnath, Ronghang Hu, Didac Suris, Chaitanya Ryali, Kalyan Vasudev Alwala, Haitham Khedr, An- drew Huang, Jie Lei, Tengyu Ma, Baishan Guo, Arpit Kalla, Markus Marks, Joseph Greer, Meng Wang, Peize Sun, Roman Rädle, Triantafyllos Afouras, Effrosyni Mavroudi, Katherine Xu, Tsung-Han Wu, Yu Zhou, Lilian...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  9. [10]

    Turgay Celik. 2009. Unsupervised Change Detection in Satellite Images Using Principal Component Analysis and 𝑘-Means Clustering.IEEE Geoscience and Remote Sensing Letters6, 4 (2009), 772–776. doi:10.1109/LGRS.2009.2025059

    work page doi:10.1109/lgrs.2009.2025059 2009

  10. [11]

    Hao Chen, Zipeng Qi, and Zhenwei Shi. 2022. Remote Sensing Image Change Detection With Transformers.IEEE Transactions on Geoscience and Remote Sensing60 (2022), 1–14. doi:10.1109/TGRS.2021.3095166

    work page doi:10.1109/tgrs.2021.3095166 2022

  11. [12]

    Hao Chen and Zhenwei Shi. 2020. A Spatial-Temporal Attention-Based Method and a New Dataset for Remote Sensing Image Change Detection.Remote Sensing 12, 10 (2020)

    2020

  12. [13]

    Hongruixuan Chen, Jian Song, Chen Wu, Bo Du, and Naoto Yokoya. 2023. Ex- change means change: An unsupervised single-temporal change detection frame- work based on intra- and inter-image patch exchange.ISPRS Journal of Pho- togrammetry and Remote Sensing206 (2023), 87–105. doi:10.1016/j.isprsjprs.2023. 11.004

    work page doi:10.1016/j.isprsjprs.2023 2023

  13. [14]

    Hongruixuan Chen, Naoto Yokoya, and Marco Chini. 2023. Fourier domain structural relationship analysis for unsupervised multimodal change detection. ISPRS Journal of Photogrammetry and Remote Sensing198 (2023), 99–114. doi:10. 1016/j.isprsjprs.2023.03.004

    2023

  14. [15]

    ZiJian Chen, YongHong Song, Yue Ma, GuoFu Li, Rui Wang, and Hao Hu. 2023. Interaction in Transformer for Change Detection in VHR Remote Sensing Images. IEEE Transactions on Geoscience and Remote Sensing61 (2023), 1–12. doi:10.1109/ TGRS.2023.3324025

    work page arXiv 2023

  15. [16]

    Schwing, Alexander Kirillov, and Rohit Girdhar

    Bowen Cheng, Ishan Misra, Alexander G. Schwing, Alexander Kirillov, and Rohit Girdhar. 2022. Masked-attention Mask Transformer for Universal Image Segmentation.CVPR

    2022

  16. [17]

    J. S. Deng, K. Wang, Y. H. Deng, and G. J. Qi. 2008. PCA-based land-use change detection and analysis using multitemporal and multisensor satel- lite data.International Journal of Remote Sensing29, 16 (2008), 4823–4838. doi:10.1080/01431160801950162

    work page doi:10.1080/01431160801950162 2008

  17. [18]

    Peijun Du, Xin Wang, Dongmei Chen, Sicong Liu, Cong Lin, and Yaping Meng

  18. [19]

    doi:10.1016/j.isprsjprs.2020.01.026

    An improved change detection approach using tri-temporal logic-verified change vector analysis.ISPRS Journal of Photogrammetry and Remote Sensing 161 (2020), 278–293. doi:10.1016/j.isprsjprs.2020.01.026

    work page doi:10.1016/j.isprsjprs.2020.01.026 2020

  19. [20]

    Yuchao Feng, Jiawei Jiang, Honghui Xu, and Jianwei Zheng. 2023. Change De- tection on Remote Sensing Images using Dual-branch Multi-level Inter-temporal Network.IEEE Transactions on Geoscience and Remote Sensing(2023), 1–1. doi:10.1109/TGRS.2023.3241257

    work page doi:10.1109/tgrs.2023.3241257 2023

  20. [21]

    Feng Gao, Xiao Wang, Yunhao Gao, Junyu Dong, and Shengke Wang. 2019. Sea Ice Change Detection in SAR Images Based on Convolutional-Wavelet Neural Networks.IEEE Geoscience and Remote Sensing Letters16, 8 (2019), 1240–1244

    2019

  21. [22]

    Gil-Yepes, Luis A

    Jose L. Gil-Yepes, Luis A. Ruiz, Jorge A. Recio, Ángel Balaguer-Beser, and Tx- omin Hermosilla. 2016. Description and validation of a new set of object- based temporal geostatistical features for land-use/land-cover change detec- tion.ISPRS Journal of Photogrammetry and Remote Sensing121 (2016), 77–91. doi:10.1016/j.isprsjprs.2016.08.010

    work page doi:10.1016/j.isprsjprs.2016.08.010 2016

  22. [23]

    Schumann, Paul D

    Laura Giustarini, Renaud Hostache, Patrick Matgen, Guy J.-P. Schumann, Paul D. Bates, and David C. Mason. 2013. A Change Detection Approach to Flood Mapping in Urban Areas Using TerraSAR-X.IEEE Transactions on Geoscience and Remote Sensing51, 4 (2013), 2417–2430

    2013

  23. [24]

    Chengxi Han, Chen Wu, Haonan Guo, Meiqi Hu, Jiepan Li, and Hongruixuan Chen. 2023. Change Guiding Network: Incorporating Change Prior to Guide Change Detection in Remote Sensing Imagery.IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing16 (2023), 8395–8407. doi:10. 1109/JSTARS.2023.3310208

    work page arXiv 2023

  24. [25]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2015. Deep Residual Learning for Image Recognition. arXiv:1512.03385 [cs.CV] https://arxiv.org/abs/ 1512.03385

    work page internal anchor Pith review arXiv 2015

  25. [26]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep Residual Learning for Image Recognition. InProc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 770–778

    2016

  26. [27]

    Thorsten Hoberg, Franz Rottensteiner, Raul Queiroz Feitosa, and Christian Heipke. 2015. Conditional Random Fields for Multitemporal and Multiscale Classification of Optical Satellite Imagery.IEEE Transactions on Geoscience and Remote Sensing53, 2 (2015), 659–673. doi:10.1109/TGRS.2014.2326886

    work page doi:10.1109/tgrs.2014.2326886 2015

  27. [28]

    Masroor Hussain, Dongmei Chen, Angela Cheng, Hui Wei, and David Stanley

  28. [29]

    doi:10.1016/j.isprsjprs.2013.03.006

    Change detection from remotely sensed images: From pixel-based to object-based approaches.ISPRS Journal of Photogrammetry and Remote Sensing 80 (2013), 91–106. doi:10.1016/j.isprsjprs.2013.03.006

    work page doi:10.1016/j.isprsjprs.2013.03.006 2013

  29. [30]

    Shunping Ji, Shiqing Wei, and Meng Lu. 2019. Fully Convolutional Networks for Multisource Building Extraction From an Open Aerial and Satellite Imagery Data Set.IEEE Transactions on Geoscience and Remote Sensing57, 1 (2019), 574–586

    2019

  30. [31]

    Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc Le, Yun-Hsuan Sung, Zhen Li, and Tom Duerig. 2021. Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision. InProc. Int. Conf. Machine Learning. (ICML), Vol. 139. 4904–4916

    2021

  31. [32]

    Kaixuan Jiang and Chen Wu. 2025. LGCANet: Local–Global and Change-Aware Network via Segment Anything Model for Remote Sensing Images Change Detection.IEEE Transactions on Geoscience and Remote Sensing63 (2025), 1–13. doi:10.1109/TGRS.2025.3582784

    work page doi:10.1109/tgrs.2025.3582784 2025

  32. [33]

    Yufeng Jiang and Yiqing Shen. 2024. M4oE: A Foundation Model for Medical Multimodal Image Segmentation with Mixture of Experts . Inproceedings of Medical Image Computing and Computer Assisted Intervention – MICCAI 2024, Vol. LNCS 15012. Springer Nature Switzerland

    2024

  33. [34]

    Kasetkasem and P.K

    T. Kasetkasem and P.K. Varshney. 2002. An image change detection algorithm based on Markov random field models.IEEE Transactions on Geoscience and Remote Sensing40, 8 (2002), 1815–1823. doi:10.1109/TGRS.2002.802498

    work page doi:10.1109/tgrs.2002.802498 2002

  34. [35]

    Segment Anything

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C. Berg, Wan-Yen Lo, Piotr Dollár, and Ross Girshick. 2023. Segment Anything.arXiv:2304.02643(2023)

    work page internal anchor Pith review arXiv 2023

  35. [36]

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. 2012. ImageNet Clas- sification with Deep Convolutional Neural Networks. InProc. Adv. Neural Inf. Process. Syst. (NIPS), Vol. 25

    2012

  36. [38]

    Bingyu Li, Haocheng Dong, Da Zhang, Zhiyuan Zhao, Junyu Gao, and Xuelong Li. 2025. Exploring Efficient Open-Vocabulary Segmentation in the Remote Sensing.arXiv preprint arXiv:2509.12040(2025)

    work page arXiv 2025

  37. [39]

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation. InICML

    2022

  38. [40]

    Kaiyu Li, Xiangyong Cao, Yupeng Deng, Chao Pang, Zepeng Xin, Deyu Meng, and Zhi Wang. 2025. DynamicEarth: How Far are We from Open-Vocabulary Change Detection?ArXivabs/2501.12931 (2025)

    work page arXiv 2025

  39. [41]

    Kaiyu Li, Xiangyong Cao, Yupeng Deng, Jiayi Song, Junmin Liu, Deyu Meng, and Zhi Wang. 2024. SemiCD-VL: Visual-language model guidance makes better semi-supervised change detector.IEEE Transactions on Geoscience and Remote Sensing(2024)

    2024

  40. [42]

    Kaiyu Li, Ruixun Liu, Xiangyong Cao, Xueru Bai, Feng Zhou, Deyu Meng, and Zhi Wang. 2025. Segearth-ov: Towards training-free open-vocabulary segmentation Seg2Change: Adapting Open-Vocabulary Semantic Segmentation Model for Remote Sensing Change Detection for remote sensing images. InProceedings of the Computer Vision and Pattern Recognition Conference. 10...

    2025

  41. [43]

    Kaiyu Li, Shengqi Zhang, Yupeng Deng, Zhi Wang, Deyu Meng, and Xiangyong Cao. 2025. SegEarth-OV3: Exploring SAM 3 for Open-Vocabulary Semantic Segmentation in Remote Sensing Images.arXiv preprint arXiv:2512.08730(2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  42. [44]

    Zhiwei Li, Huanfeng Shen, Qing Cheng, Yuhao Liu, Shucheng You, and Zongyi He. 2019. Deep learning based cloud detection for medium and high resolution remote sensing images of different sensors.ISPRS Journal of Photogrammetry and Remote Sensing150 (2019), 197–212

    2019

  43. [46]

    Zhenglai Li, Chang Tang, Lizhe Wang, and Albert Y. Zomaya. 2022. Remote Sensing Change Detection via Temporal Feature Interaction and Guided Re- finement.IEEE Transactions on Geoscience and Remote Sensing60 (2022), 1–11. doi:10.1109/TGRS.2022.3199502

    work page doi:10.1109/tgrs.2022.3199502 2022

  44. [47]

    Fan Liu, Delong Chen, Zhangqingyun Guan, Xiaocong Zhou, Jiale Zhu, Qiaolin Ye, Liyong Fu, and Jun Zhou. 2024. Remoteclip: A vision language foundation model for remote sensing.IEEE Transactions on Geoscience and Remote Sensing 62 (2024), 1–16

    2024

  45. [48]

    Mengxi Liu, Zhuoqun Chai, Haojun Deng, and Rong Liu. 2022. A CNN- Transformer Network With Multiscale Context Aggregation for Fine-Grained Cropland Change Detection.IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens. 15 (2022), 4297–4306. doi:10.1109/JSTARS.2022.3177235

    work page doi:10.1109/jstars.2022.3177235 2022

  46. [49]

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chun- yuan Li, Jianwei Yang, Hang Su, Jun Zhu, et al. 2023. Grounding dino: Marrying dino with grounded pre-training for open-set object detection.arXiv preprint arXiv:2303.05499(2023)

    work page Pith review arXiv 2023

  47. [50]

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. 2021. Swin Transformer: Hierarchical Vision Transformer using Shifted Windows . In2021 IEEE/CVF International Conference on Computer Vision (ICCV). IEEE Computer Society, Los Alamitos, CA, USA, 9992–10002. doi:10.1109/ICCV48922.2021.00986

    work page doi:10.1109/iccv48922.2021.00986 2021

  48. [51]

    Ziyuan Liu, Ruifei Zhu, Long Gao, Yuanxiu Zhou, Jingyu Ma, and Yuantao Gu. 2025. JL1-CD: A New Benchmark for Remote Sensing Change Detection and a Robust Multi-Teacher Knowledge Distillation Framework.arXiv preprint arXiv:2502.13407(2025)

    work page arXiv 2025

  49. [52]

    Prabuddh Kumar Mishra, Aman Rai, and Suresh Chand Rai. 2020. Land use and land cover change detection using geospatial techniques in the Sikkim Himalaya, India.The Egyptian Journal of Remote Sensing and Space Science23, 2 (2020), 133–143

    2020

  50. [53]

    Nielsen, Knut Conradsen, and James J

    Allan A. Nielsen, Knut Conradsen, and James J. Simpson. 1998. Multivariate Al- teration Detection (MAD) and MAF Postprocessing in Multispectral, Bitemporal Image Data: New Approaches to Change Detection Studies.Remote Sensing of Environment64, 1 (1998), 1–19. doi:10.1016/S0034-4257(97)00162-4

    work page doi:10.1016/s0034-4257(97)00162-4 1998

  51. [54]

    Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El- Nouby, et al. 2023. Dinov2: Learning robust visual features without supervision. ArXivabs/2304.07193 (2023)

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  52. [55]

    Nobuyuki Otsu. 1979. A Threshold Selection Method from Gray-Level His- tograms.IEEE Transactions on Systems, Man, and Cybernetics9, 1 (1979), 62–66

    1979

  53. [56]

    Eleonora Jonasova Parelius. 2023. A Review of Deep-Learning Methods for Change Detection in Multispectral Remote Sensing Images.Remote Sensing15, 8 (2023). doi:10.3390/rs15082092

    work page doi:10.3390/rs15082092 2023

  54. [57]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sand- hini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al

  55. [58]

    Learning transferable visual models from natural language supervision. In Proc. Int. Conf. Machine Learning. (ICML). 8748–8763

  56. [59]

    Radke, S

    R.J. Radke, S. Andra, O. Al-Kofahi, and B. Roysam. 2005. Image change detection algorithms: a systematic survey.IEEE Transactions on Image Processing14, 3 (2005), 294–307. doi:10.1109/TIP.2004.838698

    work page doi:10.1109/tip.2004.838698 2005

  57. [60]

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman Rädle, Chloe Rolland, Laura Gustafson, Eric Mintun, Junting Pan, Kalyan Vasudev Alwala, Nicolas Carion, Chao-Yuan Wu, Ross Girshick, Piotr Dollár, and Christoph Feichtenhofer. 2024. SAM 2: Segment Anything in Images and Videos.arXiv preprint arXiv:24...

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  58. [61]

    Evan Shelhamer, Jonathan Long, and Trevor Darrell. 2017. Fully Convolutional Networks for Semantic Segmentation.IEEE Transactions on Pattern Analysis and Machine Intelligence39, 4 (2017), 640–651

    2017

  59. [62]

    Yunhang Shen, Chaoyou Fu, Peixian Chen, Mengdan Zhang, Ke Li, Xing Sun, Yunsheng Wu, Shaohui Lin, and Rongrong Ji. 2024. Aligning and Prompting Everything All at Once for Universal Visual Perception.CVPR

    2024

  60. [63]

    Qian Shi, Mengxi Liu, Shengchen Li, Xiaoping Liu, Fei Wang, and Liangpei Zhang

  61. [64]

    doi:10.1109/TGRS.2021.3085870

    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(2021), 1–16. doi:10.1109/TGRS.2021.3085870

    work page doi:10.1109/tgrs.2021.3085870 2021

  62. [65]

    Wenzhong Shi, Min Zhang, Rui Zhang, Shanxiong Chen, and Zhao Zhan. 2020. Change Detection Based on Artificial Intelligence: State-of-the-Art and Chal- lenges.Remote Sensing12, 10 (2020). doi:10.3390/rs12101688

    work page doi:10.3390/rs12101688 2020

  63. [66]

    Richards

    Chunqiao Song, Bo Huang, Linghong Ke, and Keith S. Richards. 2014. Remote sensing of alpine lake water environment changes on the Tibetan Plateau and surroundings: A review.ISPRS Journal of Photogrammetry and Remote Sensing 92 (2014), 26–37

    2014

  64. [67]

    Xiaoliang Tan, Guanzhou Chen, Tong Wang, Jiaqi Wang, and Xiaodong Zhang

  65. [68]

    Segment change model (scm) for unsupervised change detection in vhr remote sensing images: a case study of buildings. InProc. IEEE Int. Geoscience And Remote Sensing Sym. (IGARSS). 8577–8580

  66. [69]

    Xiaoliang Tan, Guanzhou Chen, Xiaodong Zhang, Tong Wang, Jiaqi Wang, Kui Wang, and Tingxuan Miao. 2025. TripleS: Mitigating multi-task learning conflicts for semantic change detection in high-resolution remote sensing imagery.ISPRS Journal of Photogrammetry and Remote Sensing230 (2025), 374–401. doi:10.1016/ j.isprsjprs.2025.09.019

    2025

  67. [70]

    Xiaoliang Tan, Guanzhou Chen, Xiaodong Zhang, Tong Wang, Jiaqi Wang, Kui Wang, and Tingxuan Miao. 2025. TripleS: Mitigating multi-task learning conflicts for semantic change detection in high-resolution remote sensing imagery.ISPRS Journal of Photogrammetry and Remote Sensing230 (2025), 374–401

    2025

  68. [71]

    Xu Tang, Tianxiang Zhang, Jingjing Ma, Xiangrong Zhang, Fang Liu, and Licheng Jiao. 2023. WNet: W-Shaped Hierarchical Network for Remote-Sensing Image Change Detection.IEEE Transactions on Geoscience and Remote Sensing61 (2023), 1–14. doi:10.1109/TGRS.2023.3296383

    work page doi:10.1109/tgrs.2023.3296383 2023

  69. [72]

    Chen Wu, Bo Du, and Liangpei Zhang. 2014. Slow Feature Analysis for Change Detection in Multispectral Imagery.IEEE Transactions on Geoscience and Remote Sensing52, 5 (2014), 2858–2874. doi:10.1109/TGRS.2013.2266673

    work page doi:10.1109/tgrs.2013.2266673 2014

  70. [73]

    George Xian, Collin Homer, and Joyce Fry. 2009. Updating the 2001 National Land Cover Database land cover classification to 2006 by using Landsat imagery change detection methods.Remote Sensing of Environment113, 6 (2009), 1133–

    2009

  71. [74]

    doi:10.1016/j.rse.2009.02.004

    work page doi:10.1016/j.rse.2009.02.004 2009

  72. [75]

    Xu, Wenhan Lu, Zebo Li, Pranav Khaitan, and Valeriya Zaytseva

    Joseph Z. Xu, Wenhan Lu, Zebo Li, Pranav Khaitan, and Valeriya Zaytseva. 2019. Building Damage Detection in Satellite Imagery Using Convolutional Neural Networks.ArXivabs/1910.06444 (2019)

    work page arXiv 2019

  73. [76]

    Kunping Yang, Gui-Song Xia, Zicheng Liu, Bo Du, Wen Yang, and Marcello Pelillo. 2020. Asymmetric Siamese Networks for Semantic Change Detection. CoRRabs/2010.05687 (2020). arXiv:2010.05687

    work page arXiv 2020

  74. [77]

    Lihe Yang, Zhen Zhao, and Hengshuang Zhao. 2025. UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation.TPAMI(2025)

    2025

  75. [79]

    Zilu Ying, Zijun Tan, Yikui Zhai, Xudong Jia, Wenba Li, Junying Zeng, Angelo Genovese, Vincenzo Piuri, and Fabio Scotti. 2024. DGMA2-Net: A Difference- Guided Multiscale Aggregation Attention Network for Remote Sensing Change Detection.IEEE Transactions on Geoscience and Remote Sensing62 (2024), 1–16. doi:10.1109/TGRS.2024.3390206

    work page doi:10.1109/tgrs.2024.3390206 2024

  76. [80]

    Panli Yuan, Qingzhan Zhao, Xingbiao Zhao, Xuewen Wang, Xuefeng Long, and Yuchen Zheng. 2022. A transformer-based Siamese network and an open optical dataset for semantic change detection of remote sensing images.International Journal of Digital Earth15, 1 (2022), 1506–1525. doi:10.1080/17538947.2022. 2111470

    work page doi:10.1080/17538947.2022 2022

  77. [82]

    Tao Zhan, Qiushi Tian, Yuanyuan Zhu, Jie Lan, Qianlong Dang, and Maoguo Gong. 2025. Difference-Aware Multiscale Feature Aggregation Network for Building Change Detection.IEEE Transactions on Geoscience and Remote Sensing 63 (2025), 1–15. doi:10.1109/TGRS.2025.3560977

    work page doi:10.1109/tgrs.2025.3560977 2025

  78. [83]

    Chenxiao Zhang, Peng Yue, Deodato Tapete, Liangcun Jiang, Boyi Shangguan, Li Huang, and Guangchao Liu. 2020. A deeply supervised image fusion network for change detection in high resolution bi-temporal remote sensing images.ISPRS Journal of Photogrammetry and Remote Sensing166 (2020), 183–200

    2020

  79. [84]

    Xu Zhao, Wenchao Ding, Yongqi An, Yinglong Du, Tao Yu, Min Li, Ming Tang, and Jinqiao Wang. 2023. Fast Segment Anything.ArXiv2306.12156 (2023)

    work page arXiv 2023

  80. [85]

    Zhuo Zheng, Yanfei Zhong, Liangpei Zhang, and Stefano Ermon. 2024. Segment any change.Proc. Adv. Neural Inf. Process. Syst. (NIPS)37 (2024), 81204–81224

    2024

Showing first 80 references.