Pith. sign in

REVIEW 4 major objections 5 minor 57 references

Plug-and-Play DISep: Separating Dense Instances for Scene-to-Pixel Weakly-Supervised Change Detection in High-Resolution Remote Sensing Images

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Adding a training-time instance-separation module to weakly supervised change detection splits densely packed changed objects that are incorrectly merged, improving pixel-level accuracy across seven weak-supervision baselines and five…

desk verdict A solid plug-and-play module for reducing instance lumping in weakly supervised change detection, with consistent gains across many baselines, but the core mechanism shows threshold fragility and the evaluation lacks error bars. read the letter →

arxiv 2501.04934 v2 pith:BN5UXVUB submitted 2025-01-09 cs.CV

Zhenghui Zhao , Chen Wu , Lixiang Ru , Di Wang , Hongruixuan Chen , Cuiqun Chen This is my paper · ORCID
classification cs.CV
keywords weaklysupervisedchangedetectioninstancelumpingclassactivationmapsdenseinstancesseparationlossconnectivitysearchremotesensingplug-and-playmodule
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 proposes DISep, a plug-and-play training module for weakly supervised change detection (WSCD), where only scene-level change labels are available. Its central claim is that dense clusters of changed objects are routinely merged in pixel-level predictions, and that a three-step module—localize instance candidates via a high-pass threshold on class activation maps, group pixels into instance IDs by connectivity search, and enforce intra-instance pixel-feature consistency with a separation loss—reliably unpicks this lumping. Across three Transformer-based and four ConvNet-based baselines on the LEVIR-CD, WHU-CD, DSIFN-CD, SYSU-CD, and CDD datasets, DISep consistently improves F1 and IoU, e.g., +6.27 F1 and +6.44 IoU for TransWCD on LEVIR-CD, while adding negligible training cost and no inference cost. If correct, the result matters because accurate instance separation is what makes change quantification—counting damaged buildings, new structures, or lost vegetation—possible under cheap scene-level annotation.

What carries the argument

The module is a three-step loop around any WSCD classifier. First, Instance Localization thresholds the class activation map $C$ at a high-pass value $T_h$ to produce a binary mask of reliably changed pixels; unchanged background below $T_l$ is treated as a single unchanged instance. Second, Instance Retrieval runs an 8-neighborhood connectivity search over that mask, assigning every connected component a distinct instance ID $k$ and producing an instance identity mask $M_{id}$. Third, Instance Separation adds a separation loss $L_{\text{sep}} = l_{pc} + l_{puc} + l_{pu}$, where each term is the mean squared Euclidean distance from pixels in an instance to the instance's centroid feature computed on the last-layer feature map $F$, covering changed instances in changed images, unchanged background in changed images, and whole unchanged images. The loss is weighted by $\alpha = 0.1$ and added to the classification cross-entropy; the whole loop is iterative, so the instance IDs refresh as the features improve.

What would settle it

Take a synthetic pair of bi-temporal images with two changed rectangles that touch at a single corner, so that the CAM at threshold $T_h = 0.60$ cannot separate them into two connected components. Train a baseline plus DISep and compare the number of predicted change instances: if the module labels both rectangles as one instance and the loss then pulls their pixel features together, the predicted change map will still be merged, demonstrating that the claimed separation is limited to objects that the threshold already disjoins.

Watch

Extended reading notes

Core claim

DISep's central discovery is that instance identity, normally unavailable under scene-to-pixel supervision, can be manufactured from the same classification signals already used for change localization, and that using those manufactured IDs to cluster pixel embeddings sharpens change boundaries. The paper states that it 'achieves state-of-the-art performance by enhancing three Transformer-based and four ConvNet-based methods' on all five datasets, with consistent gains in F1, IoU, and overall accuracy for every baseline tested. The mechanism is training-time-only: once the model is trained, DISep's losses are removed and inference is unchanged, so the improvement comes without any architectural modification or added runtime.

Load-bearing premise

The success of DISep hinges on one premise: the high-pass threshold applied to the CAM actually splits densely packed changed objects into separate regions; if two objects touch in the thresholded mask they are merged into one instance ID, and the separation loss then actively pushes their pixels to look alike, deepening the very lumping the method is meant to cure.

Editorial extensions

If this is right

  • Existing WSCD pipelines can be upgraded by adding DISep's three loss terms during training, gaining roughly +2 to +8 F1 points across ConvNet and Transformer baselines without any inference-time change.
  • Accurate counting of changed objects becomes feasible under scene-level supervision, since separated instance predictions let users count damaged buildings or new structures instead of reporting merged blobs.
  • The same instance-separation mechanism transfers to fully supervised change detection, improving SNUNet by +1.26 F1 and CTD-Former by +1.32 F1 on LEVIR-CD.
  • Because the module acts purely on pixel-to-centroid clustering of the model's own features, it is architecture-agnostic and could be inserted into other dense prediction heads that suffer from object merging.

Reading between the lines

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

  • I would expect the fixed threshold $T_h = 0.60$ to become a robustness bottleneck: the paper's own ablation shows that threshold choice flips the method from its best F1 to below baseline, so an adaptive per-image or per-object threshold is a natural next test.
  • The idea of manufacturing instance IDs by connectivity search on weakly supervised CAMs and then clustering pixel embeddings is not change-detection-specific; the same loop could sharpen weakly supervised instance segmentation of buildings, roads, or other remote-sensing objects.
  • The failure of centroid-wise contrast suggests DISep's separating power comes from compactifying each instance's features rather than pushing instances apart, which implies the method should be stress-tested on scenes with extremely wide instance size disparity, where intra-instance compactness alone may be insufficient.
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 / 5 minor

Summary. The paper proposes DISep, a plug-and-play training-time module for weakly supervised change detection (WSCD). It localizes changed-instance candidate regions by thresholding class activation maps at a high-pass threshold, assigns instance IDs via 8-neighborhood connectivity search, and then applies a separation loss that clusters pixels of each instance toward that instance's feature centroid. The module is added to the training objective of existing WSCD methods without changing inference. The authors evaluate DISep on top of seven WSCD baselines (three Transformer-based and four ConvNet-based) across five datasets and report consistent F1/IoU/OA improvements, including headline gains of +6.27 F1 and +6.44 IoU for TransWCD on LEVIR-CD. They also show that DISep can improve two fully supervised change detection methods.

Significance. If the central claim holds, DISep is a genuinely useful, low-cost module for a real failure mode in WSCD: instance lumping in dense scenes. The strengths of the paper are the broad experimental coverage (seven baselines, five datasets), the consistent positive deltas in nearly all settings, the ablation study over loss weight, thresholds, loss functions, and sampling scopes, the negligible inference overhead, and the release of code. These are substantive. The concern is that the specific quantity the method claims to improve—instance-level separation and the ability to count changed objects—is never directly measured, and the core mechanism relies on a hand-set CAM threshold whose misspecification can make the method perform worse than the baseline, as shown in Table 4. With additional instance-level evaluation and a more careful treatment of threshold sensitivity, the paper would provide stronger support for its claims.

major comments (4)
  1. [Section 4.2, Tables 1–2] The motivation and title center on instance lumping and on accurately quantifying the number of changed objects, but all reported metrics are pixel-level F1, OA, and IoU. No instance-level metric is reported: for example, the precision/recall of detected changed instances, the number of connected components before and after DISep, or split/merge errors. As a result, the experiments demonstrate pixel-level accuracy gains but do not directly demonstrate that instances are actually separated or that counting improves. I recommend adding an instance-level evaluation on at least LEVIR-CD and WHU-CD (using connected components of the binary change ground truth as instance proxies), or softening the claims about instance separation and quantification.
  2. [Section 3.3, Eq. (2), Table 4] The instance retrieval step is load-bearing: it performs connectivity search on the high-pass thresholded mask Mc. If two changed objects touch after thresholding, they receive the same instance ID, and the separation loss in Eq. (8) then pulls their pixels toward one centroid, potentially reinforcing the lumping the method is designed to fix. Table 4 shows that this is not merely hypothetical: setting both Th and Tl to 0.45 drops F1 from 60.08 to 58.72, below the baseline, and setting Th=0.50/Tl=0.40 also underperforms the selected configuration. The chosen Th=0.60/Tl=0.40 is ablated only on LEVIR-CD and then applied uniformly to all five datasets. Given the different instance densities across datasets (Fig. 2), the authors should either justify the transferability of the fixed thresholds, provide per-dataset sensitivity analyses, or design an adaptive threshold scheme.
  3. [Section 3.4, Eq. (8), Section 4.3.4] The separation loss only enforces intra-instance compactness; it contains no term that pushes different instances apart. Any separation achieved therefore comes from the connectivity structure already present in the high-passed CAM. The paper's own ablation in Table 5 shows that adding a centroid-to-centroid contrast term reduces F1 by 2.10%, which raises the question of whether an explicit inter-instance separation term is viable. The claim in Section 3.4 that close instance centroids are 'not a concern' is unsupported by any quantitative analysis. I would like the authors to analyze whether the method can ever correct an instance merge that survives the high-pass threshold, and to provide evidence—for example, feature-distance statistics before and after DISep—that the loss genuinely separates merged components.
  4. [Abstract and Section 4.2] The abstract and Section 4.2 claim 'state-of-the-art performance', but Tables 1–2 only compare each of the seven methods with and without +DISep. There is no comparison to external WSCD methods that are not used as backbones; for instance, CS-WSCDNet is discussed in Section 2.1 but does not appear in any experiment. Without such comparisons, the SOTA claim is not supported by the presented evidence. I recommend either adding external baseline results on at least one dataset per table, or revising the wording to 'consistently improves existing WSCD methods'.
minor comments (5)
  1. [Section 4.2.1] The text reports WCDNet's WHU-CD IoU improvement as +8.24%, but Table 1 shows +6.24% (26.34 vs. 20.10); the text and table should be aligned.
  2. [Eq. (1)] The notation C = CAM_i / Max(CAM_i) is unclear because the subscript i appears only on the numerator and the sum in the definition of CAM is over the channel dimension; please define the normalization over the spatial extent of the activation map and use consistent indexing.
  3. [Table 4] The layout of Table 4 is difficult to read because Th and Tl values are paired in a single column without grouping; consider using separate columns with clear row grouping or spacing to show the two-dimensional threshold grid.
  4. [References] The reference list contains the same paper by Shi et al. twice, once as 'Shi et al., 2021' and once as 'Shi et al., 2022', with the same DOI; this duplicate should be consolidated.
  5. [Section 5.1, Figures 8–9] The t-SNE visualizations support the claim of improved feature grouping but are qualitative; adding a simple quantitative cluster metric (e.g., average intra-instance vs. inter-instance feature distance, or silhouette score) would make the evidence more convincing.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DISep's instance-separation loss is a self-training auxiliary objective, and the paper's central improvement claim is evaluated against external pixel-level benchmarks across seven baselines and five datasets.

full rationale

The paper does not derive its headline result from its own inputs by construction. The separation loss (Eq. 8) is defined on instance masks obtained by thresholding the model's own CAM (Eq. 2) and connectivity search, so the auxiliary objective is a pseudo-label self-training loop rather than a fitted parameter renamed as a prediction. The central claim—that DISep reduces instance lumping and improves pixel-level F1/IoU—is measured on held-out ground-truth labels, not on the instance masks themselves. The thresholds Th and Tl are hyperparameters chosen by ablation (Table 4), not quantities predicted by the method. The paper self-cites TransWCD and FCD-GAN as baselines, but these citations are not load-bearing premises: TransWCD is one of seven baselines, and the other five baselines come from independent groups. No uniqueness theorem from the authors is invoked to forbid alternatives, and no known result is merely renamed. The main vulnerability—that merged CAM regions receive one instance ID and the loss then reinforces that grouping—is an empirical robustness concern, not a circular derivation, because the final evaluation is independent of the pseudo-instance assignments.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical or conceptual entities; it combines known building blocks (CAMs, connected components, clustering loss) into a new training module. The main load-bearing elements are the threshold choices and the assumption that CAM peaks separate dense instances. These are recorded as free parameters and domain assumptions above. The invented_entities list is empty because DISep is a loss and training procedure, not a new object, force, or conserved quantity.

free parameters (5)
  • High-pass threshold Th = 0.60
    Used to build the changed instance localization mask from CAMs (Eq. 2). Chosen by ablation on LEVIR-CD (Table 4); the paper shows performance degrades when set to 0.45 or other values.
  • Low-pass threshold Tl = 0.40
    Used to define the unchanged background instance mask (Eq. 3). Selected together with Th via ablation on LEVIR-CD (Table 4).
  • Separation loss weight alpha = 0.1
    Controls the contribution of Lsep to the total loss (Eq. 13). Tuned by experiment on WHU-CD, LEVIR-CD, and DSIFN-CD (Figure 7).
  • CAM score threshold for final pseudo-labels = 0.45
    Used to differentiate changed and unchanged regions for the final prediction. Mentioned in Section 4.1.4 as a preset value; the paper notes that setting both Th and Tl to 0.45 degrades performance.
  • Integration iteration for Lsep = after 200 iterations
    The separation loss is applied only after the first 200 training iterations, described in Section 4.1.4. This is a hyperparameter but not ablated.
assumptions (4)
  • domain assumption Class activation maps under scene-level supervision can localize changed instances, and high-pass thresholding can separate dense instances.
    The entire DISep pipeline depends on the existence of CAM peaks for individual changed objects even when predictions at a lower threshold merge them. Cited prior work (Zhou et al., 2018) supports peak response localization, but the effectiveness on dense remote sensing instances is an assumption tested only indirectly.
  • domain assumption Connectivity search on the binary localization mask correctly groups pixels belonging to the same physical changed object.
    The instance retrieval step (Section 3.3) assumes that isolated connected components in the thresholded mask correspond one-to-one to real instances. If two different objects touch, they are merged into one ID and the separation loss reinforces the error.
  • domain assumption Pulling intra-instance pixel features toward the instance centroid improves pixel-level change detection without harming changed/unchanged discrimination.
    This is the core training signal behind Lsep (Eqs. 7-12). The paper provides ablation evidence in Table 6, but the assumption that feature consistency within an instance is beneficial is not derived from first principles.
  • ad hoc to paper The selected hyperparameters (Th=0.60, Tl=0.40, alpha=0.1) generalize across datasets.
    These values are tuned on LEVIR-CD and then applied unchanged to WHU-CD, DSIFN-CD, SYSU-CD, and CDD. The generalization is asserted rather than demonstrated through per-dataset tuning or robustness analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Plug-and-Play DISep: Separating Dense Instances for Scene-to-Pixel Weakly-Supervised Change Detection in High-Resolution Remote Sensing Images." pith.science (2026). https://pith.science/paper/BN5UXVUB

@misc{pith2026250104934,
  author       = {Pith},
  title        = {Pith review of: Plug-and-Play DISep: Separating Dense Instances for Scene-to-Pixel Weakly-Supervised Change Detection in High-Resolution Remote Sensing Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BN5UXVUB}},
  note         = {Machine review of arXiv:2501.04934}
}
read the original abstract

Existing Weakly-Supervised Change Detection (WSCD) methods often encounter the problem of "instance lumping" under scene-level supervision, particularly in scenarios with a dense distribution of changed instances (i.e., changed objects). In these scenarios, unchanged pixels between changed instances are also mistakenly identified as changed, causing multiple changes to be mistakenly viewed as one. In practical applications, this issue prevents the accurate quantification of the number of changes. To address this issue, we propose a Dense Instance Separation (DISep) method as a plug-and-play solution, refining pixel features from a unified instance perspective under scene-level supervision. Specifically, our DISep comprises a three-step iterative training process: 1) Instance Localization: We locate instance candidate regions for changed pixels using high-pass class activation maps. 2) Instance Retrieval: We identify and group these changed pixels into different instance IDs through connectivity searching. Then, based on the assigned instance IDs, we extract corresponding pixel-level features on a per-instance basis. 3) Instance Separation: We introduce a separation loss to enforce intra-instance pixel consistency in the embedding space, thereby ensuring separable instance feature representations. The proposed DISep adds only minimal training cost and no inference cost. It can be seamlessly integrated to enhance existing WSCD methods. We achieve state-of-the-art performance by enhancing {three Transformer-based and four ConvNet-based methods} on the LEVIR-CD, WHU-CD, DSIFN-CD, SYSU-CD, and CDD datasets. Additionally, our DISep can be used to improve fully-supervised change detection methods. Code is available at https://github.com/zhenghuizhao/Plug-and-Play-DISep-for-Change-Detection.

Figures

Figures reproduced from arXiv: 2501.04934 by the authors.

Figure 1
Figure 1. Motivation for our DISep. We show the change predictions in dense [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Prevalence of dense instance distribution in change detection. We present statistics on the instance distribution within the WHU-CD and LEVIR-CD [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of our DISep. First, we obtain instance localization masks from the CAM using a high-pass threshold. Then, we implement instance retrieval [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Three cases of changed instance retrieval. The square with a red [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Examples of the WHU-CD, LEVIR-CD, DSIFN-CD, SYSU-CD, and [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Qualitative improvements of our DISep. For clarity, in the predictions, [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Weight factor of Lsep. We vary the weight factor of our separation loss on the overall loss, on top of the WCDNet and TransWCD baselines. F1 score (%) is reported on the WHU-CD, LEVIR-CD, and DSIFN-CD datasets. 4.3.2. Weight Factor of Lsep We investigate the sensitivit…
Figure 8
Figure 8. Figure 8: Evolution of instance-wise feature distribution. Notable enhancements are observed in scenarios of di [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Detailed evolution in distribution of instance-wise pixel features. [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 47 canonical work pages

  1. [1]

    , author Timofte, R

    author Andermatt, P. , author Timofte, R. , year 2020 . title A weakly supervised convolutional network for change segmentation and classification , in: booktitle Proceedings of the Asian Conference on Computer Vision

    work page 2020

  2. [2]

    , author Wei, C

    author Cao, K. , author Wei, C. , author Gaidon, A. , author Arechiga, N. , author Ma, T. , year 2019 . title Learning imbalanced datasets with label-distribution-aware margin loss , in: booktitle Advances in Neural Information Processing Systems , pp. pages 1565--1576

    work page 2019

  3. [3]

    , author Huang, X

    author Cao, Y. , author Huang, X. , author Weng, Q. , year 2023 . title A multi-scale weakly supervised learning method with adaptive online noise correction for high-resolution change detection of built-up areas . journal Remote Sensing of Environment volume 286 , pages 113471

    work page 2023

  4. [4]

    , author Shi, Z

    author Chen, H. , author Shi, Z. , year 2020 . title A spatial-temporal attention-based method and a new dataset for remote sensing image change detection . journal Remote Sensing volume 12 , pages 1662

    work page 2020

  5. [5]

    , author Song, J

    author Chen, H. , author Song, J. , author Han, C. , author Xia, J. , author Yokoya, N. , year 2024 . title Changemamba: Remote sensing change detection with spatio-temporal state space model . journal IEEE Transactions on Geoscience and Remote Sensing , pages 1--1 :10.1109/TGRS.2024.3417253

    arXiv 2024

  6. [6]

    , author Song, J

    author Chen, H. , author Song, J. , author Wu, C. , author Du, B. , author Yokoya, N. , year 2023 . title Exchange means change: An unsupervised single-temporal change detection framework based on intra-and inter-image patch exchange . journal ISPRS Journal of Photogrammetry and Remote Sensing volume 206 , pages 87--105

    work page 2023

  7. [7]

    , author Le Saux, B

    author Daudt, R.C. , author Le Saux, B. , author Boulch, A. , author Gousseau, Y. , year 2019 . title Guided anisotropic diffusion and iterative learning for weakly supervised change detection , in: booktitle Computer Vision and Pattern Recognition Workshops

    work page 2019

  8. [8]

    , author Le Saux, B

    author Daudt, R.C. , author Le Saux, B. , author Boulch, A. , author Gousseau, Y. , year 2023 . title Weakly supervised change detection using guided anisotropic diffusion . journal Machine Learning volume 112 , pages 2211--2237

    work page 2023

Show all 57 references
  1. [9]

    , author Fu, Z

    author Du, Y. , author Fu, Z. , author Liu, Q. , author Wang, Y. , year 2022 . title Weakly supervised semantic segmentation by pixel-to-prototype contrast , in: booktitle Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. pages 4320--4329

    2022

  2. [10]

    , author Krishnan, D

    author Elsayed, G. , author Krishnan, D. , author Mobahi, H. , author Regan, K. , author Bengio, S. , year 2018 . title Large margin deep networks for classification , in: booktitle Advances in Neural Information Processing Systems , pp. pages 842--852

    2018

  3. [11]

    , author Li, K

    author Fang, S. , author Li, K. , author Shao, J. , author Li, Z. , year 2022 . title Snunet-cd: A densely connected siamese network for change detection of very high-resolution images . journal IEEE Geoscience and Remote Sensing Letters volume 19 . note Art. no. 8007805

    2022

  4. [12]

    , author Javed, A

    author Fatima, N. , author Javed, A. , year 2021 . title Assessment of land use land cover change detection using geospatial techniques in southeast rajasthan . journal Journal of Geoscience and Environment Protection volume 9 , pages 18 . http://www.scirp.org/journal/PaperDow...

    2021

  5. [13]

    , author Zhang, J

    author Guo, Q. , author Zhang, J. , author Zhang, Y. , year 2021 . title Multitemporal images change detection based on ammf and spectral constraint strategy . journal IEEE Transactions on Geoscience and Remote Sensing volume 59 , pages 3444--3457 . :10.1109/TGRS.2020.3008016

    2021

  6. [14]

    , author Fedorov, A

    author Hjelm, R.D. , author Fedorov, A. , author Lavoie-Marchildon, S. , author Grewal, K. , author Bachman, P. , author Trischler, A. , author Bengio, Y. , year 2019 . title Learning deep representations by mutual information estimation and maximization , in: booktitle ICLR

    2019

  7. [15]

    , author Wang, R

    author Huang, R. , author Wang, R. , author Guo, Q. , author Wei, J. , author Zhang, Y. , author Fan, W. , author Liu, Y. , year 2023 . title Background-mixed augmentation for weakly supervised change detection . journal AAAI

    2023

  8. [16]

    , author Wei, S

    author Ji, S. , author Wei, S. , author Lu, M. , year 2018 . title Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set . journal IEEE Transactions on Geoscience and Remote Sensing volume 57 , pages 574--586

    2018

  9. [17]

    , author Karatsiolis, S

    author Kalita, I. , author Karatsiolis, S. , author Kamilaris, A. , year 2021 . title Land use change detection using deep siamese neural networks and weakly supervised learning , in: booktitle Computer Analysis of Images and Patterns: 19th International Conference, CAIP 2021,...

    2021

  10. [18]

    , author He, X

    author Khan, S. , author He, X. , author Porikli, F. , author Bennamoun, M. , author Sohel, F. , author Togneri, R. , year 2017 . title Learning deep structured network for weakly supervised change detection , in: booktitle Proceedings of the 26th International Joint Conferenc...

    2017

  11. [19]

    , author Yoo, Y

    author Kim, B. , author Yoo, Y. , author Rhee, C.E. , author Kim, J. , year 2022 . title Beyond semantic to instance segmentation: Weakly-supervised instance segmentation via semantic knowledge transfer and self-refinement , in: booktitle CVPR , pp. pages 4268--4277

    2022

  12. [20]

    , author Mintun, E

    author Kirillov, A. , author Mintun, E. , author Ravi, N. , author Mao, H. , author Rolland, C. , author Gustafson, L. , author Xiao, T. , author Whitehead, S. , author Berg, A.C. , author Lo, W.Y. , et al., year 2023 . title Segment anything . journal arXiv preprint arXiv:2304.02643

    2023 arXiv

  13. [21]

    a henb \

    author Kr \"a henb \"u hl, P. , author Koltun, V. , year 2011 . title Efficient inference in fully connected crfs with gaussian edge potentials . journal Advances in neural information processing systems volume 24

    2011

  14. [22]

    , author Lee, K

    author Lee, H. , author Lee, K. , author Kim, J.H. , author Na, Y. , author Park, J. , author Choi, J.P. , author Hwang, J.Y. , year 2021 . title Local similarity siamese network for urban land change detection on remote sensing images . journal IEEE Journal of Selected Topics...

    2021

  15. [23]

    , author Li, X

    author Li, X. , author Li, X. , author Zhang, L. , author Cheng, G. , author Shi, J. , author Lin, Z. , author Tan, S. , author Tong, Y. , year 2020 . title Improving semantic segmentation via decoupled body and edge supervision , in: booktitle Computer Vision--ECCV 2020: 16th...

    2020

  16. [24]

    , author Tang, C

    author Li, Z. , author Tang, C. , author Liu, X. , author Li, C. , author Li, X. , year 2024 . title Ms-former: Memory-supported transformer for weakly supervised change detection with patch-level annotations . journal IEEE Transactions on Geoscience and Remote Sensing

    2024

  17. [25]

    , author Bober, M

    author Liu, D. , author Bober, M. , author Kittler, J. , year 2019 . title Visual semantic information pursuit: A survey . journal IEEE transactions on pattern analysis and machine intelligence volume 43 , pages 1404--1422

    2019

  18. [26]

    , author Duan, J

    author Ma, J. , author Duan, J. , author Tang, X. , author Zhang, X. , author Jiao, L. , year 2023 . title Eatder: Edge-assisted adaptive transformer detector for remote sensing change detection . journal IEEE Transactions on Geoscience and Remote Sensing

    2023

  19. [27]

    , author Hinton, G

    author Maaten, L.v.d. , author Hinton, G. , year 2008 . title Visualizing data using t-sne . journal Journal of machine learning research volume 9 , pages 2579--2605

    2008

  20. [28]

    , author van der Pol, E

    author Mettes, P. , author van der Pol, E. , author Snoek, C. , year 2019 . title Hyperspherical prototype networks , in: booktitle NeurIPS

    2019

  21. [29]

    , year 1980

    author Mitchell, T.M. , year 1980 . title The need for biases in learning generalizations . type Technical Report . Department of Computer Science, Laboratory for Computer Science Research

    1980

  22. [30]

    , author Ocal, O

    author Nar, K. , author Ocal, O. , author Sastry, S.S. , author Ramchandran, K. , year 2019 . title Cross-entropy loss and low-rank features have responsibility for adversarial examples . journal arXiv preprint arXiv:1901.08360

    2019 arXiv

  23. [31]

    , author Bottou, L

    author Oquab, M. , author Bottou, L. , author Laptev, I. , author Sivic, J. , year 2015 . title Is object localization for free? - weakly-supervised learning with convolutional neural networks , in: booktitle CVPR , pp. pages 685--694

    2015

  24. [32]

    , author Arbelaez, P

    author Pont-Tuset, J. , author Arbelaez, P. , author Barron, J.T. , author Marques, F. , author Malik, J. , year 2017 . title Multiscale combinatorial grouping for image segmentation and object proposal generation . journal IEEE Transactions on Pattern Analysis and Machine Int...

    2017

  25. [33]

    , author Kalenichenko, D

    author Schroff, F. , author Kalenichenko, D. , author Philbin, J. , year 2015 . title Facenet: A unified embedding for face recognition and clustering , in: booktitle CVPR

    2015

  26. [34]

    , author Liu, M

    author Shi, Q. , author Liu, M. , author Li, S. , author Liu, X. , author Wang, F. , author Zhang, L. , year 2021 . title A deeply supervised attention metric-based network and an open aerial image dataset for remote sensing change detection . journal IEEE transactions on geos...

    2021

  27. [35]

    , author Liu, M

    author Shi, Q. , author Liu, M. , author Li, S. , author Liu, X. , author Wang, F. , author Zhang, L. , year 2022 . title A deeply supervised attention metric-based network and an open aerial image dataset for remote sensing change detection . journal IEEE Transactions on Geos...

    2022

  28. [36]

    , author Zisserman, A

    author Simonyan, K. , author Zisserman, A. , year 2014 . title Very deep convolutional networks for large-scale image recognition . journal arXiv preprint arXiv:1409.1556

    2014 arXiv

  29. [37]

    , author Sarkar, C

    author Singh, B. , author Sarkar, C. , year 2020 . title Monitoring urban growth and detection of land use/ land cover change in silchar city, assam and balurghat city, west bengal . journal International Journal of Innovative Technology and Exploring Engineering volume 9 . :1...

    2020

  30. [38]

    , year 2016

    author Sohn, K. , year 2016 . title Improved deep metric learning with multi-class n-pair loss objective , in: booktitle NeurIPS

    2016

  31. [39]

    , author Huang, B

    author Song, C. , author Huang, B. , author Ke, L. , author Richards, K.S. , year 2014 . title Remote sensing of alpine lake water environment changes on the tibetan plateau and surroundings: A review . journal ISPRS Journal of Photogrammetry and Remote Sensing volume 92 , pag...

    2014

  32. [40]

    , author McCallum, A

    author Sutton, C. , author McCallum, A. , year 2010 . title An introduction to conditional random fields . journal ar5iv https://ar5iv.labs.arxiv.org/html/1011.4088

    2010 arXiv

  33. [41]

    , author Mignotte, M

    author Touati, R. , author Mignotte, M. , author Dahmane, M. , year 2020 . title Anomaly feature learning for unsupervised change detection in heterogeneous images: A deep sparse residual model . journal IEEE Journal of Selected Topics in Applied Earth Observations and Remote ...

    2020

  34. [42]

    , author Zhong, Y

    author Wang, J. , author Zhong, Y. , author Zhang, L. , year 2023 a. title Change detection based on supervised contrastive learning for high-resolution remote sensing imagery . journal IEEE Transactions on Geoscience and Remote Sensing volume 61 , pages 1--16 . :10.1109/TGRS....

    2023

  35. [43]

    , author Zhang, M

    author Wang, L. , author Zhang, M. , author Shi, W. , year 2023 b. title Cs-wscdnet: Class activation mapping and segment anything model-based framework for weakly supervised change detection . journal IEEE Transactions on Geoscience and Remote Sensing

    2023

  36. [44]

    , author Du, B

    author Wu, C. , author Du, B. , author Zhang, L. , year 2023 . title Fully convolutional change detection framework with generative adversarial network for unsupervised, weakly supervised and regional supervised change detection . journal IEEE Transactions on Pattern Analysis ...

    2023

  37. [45]

    , author Homer, C

    author Xian, G. , author Homer, C. , year 2010 . title Updating the 2001 national land cover database impervious surface products to 2006 using landsat imagery change detection methods . journal Remote sensing of environment volume 114 , pages 1676--1686

    2010

  38. [46]

    , author Lu, W

    author Xu, J.Z. , author Lu, W. , author Li, Z. , author Khaitan, P. , author Zaytseva, V. , year 2019 . title Building damage detection in satellite imagery using convolutional neural networks . journal arXiv preprint arXiv:1910.06444

    2019 arXiv

  39. [47]

    , author Ouyang, W

    author Xu, L. , author Ouyang, W. , author Bennamoun, M. , author Boussaid, F. , author Sohel, F. , author Xu, D. , year 2021 . title Leveraging auxiliary tasks with affinity learning for weakly supervised semantic segmentation , in: booktitle Proceedings of the IEEE/CVF inter...

    2021

  40. [48]

    , author Fu, K

    author Zhan, Y. , author Fu, K. , author Yan, M. , author Sun, X. , author Wang, H. , author Qiu, X. , year 2017 . title Change detection based on deep siamese convolutional network for optical aerial images . journal IEEE Geoscience and Remote Sensing Letters volume 14 , page...

    2017

  41. [49]

    , author Wang, L

    author Zhang, C. , author Wang, L. , author Cheng, S. , author Li, Y. , year 2022 a. title Swinsunet: Pure transformer network for remote sensing image change detection . journal IEEE Transactions on Geoscience and Remote Sensing volume 60 . :10.1109/TGRS.2022.3160007

    2022

  42. [50]

    , author Yue, P

    author Zhang, C. , author Yue, P. , author Tapete, D. , author Jiang, L. , author Shangguan, B. , author Huang, L. , author Liu, G. , year 2020 . title A deeply supervised image fusion network for change detection in high resolution bi-temporal remote sensing images . journal ...

    2020

  43. [51]

    , author Shao, Z

    author Zhang, J. , author Shao, Z. , author Ding, Q. , author Huang, X. , author Wang, Y. , author Zhou, X. , author Li, D. , year 2023 a. title Aernet: An attention-guided edge refinement network and a dataset for remote sensing building change detection . journal IEEE Transa...

    2023

  44. [52]

    , author Zhao, X

    author Zhang, K. , author Zhao, X. , author Zhang, F. , author Ding, L. , author Sun, J. , author Bruzzone, L. , year 2023 b. title Relation changes matter: Cross-temporal difference transformer for change detection in remote sensing images . journal IEEE Transactions on Geosc...

    2023

  45. [53]

    , author Li, W

    author Zhang, Y. , author Li, W. , author Wang, Y. , author Wang, Z. , author Li, H. , year 2022 b. title Beyond classifiers: Remote sensing change detection with metric learning . journal Remote Sensing volume 14 , pages 4478

    2022

  46. [54]

    , author Ru, L

    author Zhao, Z. , author Ru, L. , author Wu, C. , year 2023 . title Exploring effective priors and efficient models for weakly-supervised change detection . journal arXiv preprint arXiv:2307.10853

    2023 arXiv

  47. [55]

    , author Zhong, Y

    author Zheng, Z. , author Zhong, Y. , author Wang, J. , author Ma, A. , author Zhang, L. , year 2021 . title Building damage assessment for rapid disaster response with a deep object-based semantic change detection framework: From natural disasters to man-made disasters . jour...

    2021

  48. [56]

    , author Khosla, A

    author Zhou, B. , author Khosla, A. , author Lapedriza, A. , author Oliva, A. , author Torralba, A. , year 2016 . title Learning deep features for discriminative localization , in: booktitle CVPR , pp. pages 2921--2929

    2016

  49. [57]

    , author Zhu, Y

    author Zhou, Y. , author Zhu, Y. , author Ye, Q. , author Qiu, Q. , author Jiao, J. , year 2018 . title Weakly supervised instance segmentation using class peak response , in: booktitle Proceedings of the IEEE conference on computer vision and pattern recognition , pp. pages 3...

    2018

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.