Pith. sign in

REVIEW 4 major objections 5 minor 57 references

ViewDelta: Scaling Scene Change Detection through Text-Conditioning

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read ViewDelta claims that a single text-conditioned model, trained jointly on several scene change detection datasets, performs as well as or better than models trained per dataset, and that text conditioning resolves the labeling conflicts…

desk verdict Strong dataset and sensible framework; the headline generalization numbers hinge on a split-provenance detail the paper never documents. read the letter →

arxiv 2412.07612 v3 pith:35J6JPDL submitted 2024-12-10 cs.CV

Subin Varghese , Joshua Gao , Vedhus Hoskere This is my paper · ORCID
classification cs.CV
keywords scenechangedetectiontext-conditionedsegmentationvision-languagemodelsjointtrainingsyntheticdatasetviewpointinvarianceCSeggeneralization
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

ViewDelta is a framework for scene change detection that lets a user define, in natural language, which changes count as relevant before the model looks at an image pair. The paper argues that this text conditioning dissolves the labeling conflicts that have kept scene change detection models from training jointly on multiple datasets, because a prompt such as "vehicle" or "all observable differences" tells the model which of the dataset's labels apply. To support the argument, the authors build CSeg, a synthetic dataset of over 500,000 image pairs with more than 300,000 unique prompts, and train a single ViewDelta model jointly on CSeg, PSCD, SYSU-CD, VL-CMU-CD, and unaligned variants of the latter two. They report that this one model is competitive with or better than dataset-specific baselines, including on unaligned image pairs with viewpoint change. If the results hold, text conditioning is a practical route to a single generalizable change detector rather than a separate model per benchmark.

What carries the argument

The machinery is a prompt-conditioned transformer: a frozen SigLip text encoder turns the prompt into tokens, a frozen Dinov2 vision transformer embeds the before and after images, and these tokens are concatenated with learnable segmentation query tokens and passed through a ViT backbone. The segmentation head reads only the segmentation query tokens, deliberately avoiding operations such as difference or concatenation of pixel-aligned features that assume the two views are spatially aligned. The other load-bearing piece is the CSeg generation procedure, which uses large vision-language model class proposals and open-vocabulary segmentation masks, inpainting to create changes, affine transforms for viewpoint variation, and "red herring" masks so the model learns to ignore inpainting artifacts and follow the text.

What would settle it

Inspect the Diff-1 and Diff-2 test splits of PSCD and VL-CMU-CD for any image whose original or perturbed version appears in the training splits; if an overlap exists, the F1 scores in Tables 4 and 5 are inflated by train/test leakage and the viewpoint-robustness claim weakens.

Watch

Extended reading notes

Core claim

The paper's central claim is that scene change detection can be made generalizable by conditioning the binary change mask on a text prompt that specifies the relevant change, and that this conditioning is strong enough to permit one ViewDelta model to be trained jointly across datasets whose labeling conventions contradict each other. On CSeg, the general model reaches 83.80 IoU; on PSCD multi-class scene change detection it reaches 51.2 IoU without dataset-specific training, within 4.3 IoU of its own fine-tuned version; on SYSU-CD it reaches 67.05 IoU; and on the unaligned Diff-1/Diff-2 variants of PSCD and VL-CMU-CD it stays close to its aligned performance. The authors interpret these numbers as showing that text prompts resolve the ambiguity of what counts as change, that avoiding spatial alignment assumptions in the segmentation head is what preserves performance under viewpoint changes, and that the general model is a good starting point for further fine-tuning.

Load-bearing premise

The results on unaligned benchmarks rest on the unstated assumption that the Diff-1 and Diff-2 variants of PSCD and VL-CMU-CD used for training were built only from the original training splits, so no test image or a perturbed version of one was seen during training.

Editorial extensions

If this is right

  • A single ViewDelta model can serve street-view, satellite, and multi-view change-detection tasks at inference time, with the user's text deciding what counts as a change.
  • Future scene change detection datasets can be pooled without manually reconciling their label definitions, because the prompt supplies the missing context for each training example.
  • The model keeps most of its accuracy when the input pair is unaligned, indicating that real-world change detection from casually captured, non-registered photos is within reach.
  • Fine-tuning the jointly trained model gives a consistent boost over training per dataset, so text-conditioned joint pretraining is a reusable starting point for specialized change detectors.

Reading between the lines

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

  • An untested implication is that ViewDelta inherits an open vocabulary from its frozen text encoder, so zero-shot evaluation on prompts like "damage" or "flooding" over unlabeled image pairs would reveal whether the generalization claim extends beyond the benchmark class lists.
  • The CSeg recipe of vision-language class proposals, segmentation masks, inpainting, and red herring masks could be reused to produce text-conditioned change data in domains such as medical imaging or aerial surveying, where no such dataset currently exists.
  • Because the segmentation head never aligns the two image streams, the same architecture may extend to video frames with large camera motion, a setting where most change detectors fail; this is an inference, not something the paper evaluates.
  • The large margin over the Gemini baseline on CSeg may partly reflect that the synthetic prompts are near the training distribution; a harder test would be human-authored prompts on real image pairs with viewpoint differences.
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 presents ViewDelta, a text-conditioned scene change detection (SCD) framework that takes two images and a natural-language prompt and outputs a binary change mask. To enable this, the authors release CSeg, a large synthetic dataset with over 500K image pairs and 300K unique prompts. ViewDelta is trained jointly on CSeg, SYSU-CD, PSCD, VL-CMU-CD, and the unaligned Diff-1/Diff-2 variants of PSCD and VL-CMU-CD. Experiments report competitive or superior performance against dataset-specific baselines, particularly on viewpoint-challenging variants, plus ablations isolating the contribution of frozen image/text embeddings and segmentation query tokens.

Significance. If the reported results are reliable, ViewDelta demonstrates a promising direction: using text prompts to disambiguate change relevance across datasets with conflicting labeling conventions, enabling a single model to work across domains and view changes. The CSeg dataset is a substantial resource. The paper includes useful ablations and a reproducible architecture. However, the headline viewpoint-robustness numbers (Tables 4–5) rest on an unverified data-provenance assumption about the unaligned variants, and the novelty claim is weakened by unaddressed prior work on text-conditioned change detection. These issues need to be resolved before the contribution can be fully assessed.

major comments (4)
  1. [Section 3.6 and Section 5.1, Tables 4 and 5] The paper trains ViewDelta on the Diff-1 and Diff-2 unaligned variants of PSCD and VL-CMU-CD and evaluates on the same variants, but it never specifies whether those variants were created solely from the original training splits. If Lin et al. [28] generated the variants from the full datasets including test images, then test-time images or near-duplicates thereof would have been seen during training, inflating the reported F1 scores (e.g., 63.6 vs. 28.4 for Dinov2 RSCD on PSCD Diff-2) and invalidating the claim of robustness to viewpoint changes. The manuscript must state and substantiate the exact split provenance, or re-run the evaluation using variants derived only from training splits.
  2. [Section 1 (Introduction) and Section 2 (Related Work)] The paper claims as a contribution 'a novel text prompt conditioned change detection task,' but prior work already exists in this area—notably Sachdeva and Zisserman's 'The Change You Want to See' [39] and its 3D extension [40], which introduce language-conditioned change detection together with a synthetic dataset (COCO-Inpainted). The authors cite these works only as synthetic dataset generation and do not compare against or discuss them as text-conditioned methods. This weakens the novelty claim; the authors should position ViewDelta relative to these methods and include empirical comparisons on a common benchmark (e.g., CSeg or COCO-Inpainted).
  3. [Section 5.1, Tables 2–5] ViewDelta results are reported as point estimates without error bars or repeated-run variance, while the Gemini baseline in Table 1 includes 95% confidence intervals. On SYSU-CD (Table 3), the differences between ViewDelta and state-of-the-art baselines are within a few IoU points (e.g., 70.09 vs. 71.10 for MambaBCD; 67.05 vs. 66.79 for TransUNetCD), so without variance estimates it is unclear whether these differences are meaningful. The authors should report confidence intervals or multiple seeds, at least for the main comparisons.
  4. [Section 5.1, Tables 3 and 5] The evaluation prompts for SYSU-CD and VL-CMU-CD are long, hand-crafted lists of classes or descriptions (e.g., 'urban development, suburban expansion, pre-construction groundwork, vegetation alteration, road widening, and coastal construction'). The manuscript provides no sensitivity analysis or protocol for selecting these prompts, leaving open the possibility that they were chosen to maximize performance. The authors should either use a validation set to select prompts or report results across multiple reasonable prompt variants; otherwise the fairness of these comparisons cannot be assessed.
minor comments (5)
  1. [Section 4.3] The CSeg statistics are inconsistent: the test set contains 7,326 unique prompts, yet the text reports 35,271 unique prompts not seen in train. Please clarify the counts.
  2. [Abstract] The dataset URL appears as 'github.io/viewdelta/' in the abstract and as 'https://joshuakgao.github.io/viewdelta/' in the contributions; the correct URL should be used consistently.
  3. [Table 1] ViewDelta is evaluated on the full CSeg test set, while Gemini 2.5 Pro is evaluated on a 2000-sample subset; the sampling procedure should be described and the uneven comparison noted in the table caption.
  4. [Table 4] ViewDelta's fine-tuned F1 of 63.1 on aligned PSCD is far above the previous best of 44.2 (Dinov2 RSCD); the paper does not analyze this large gap, and a discussion of possible reasons would strengthen the work.
  5. [Section 3.2] The sentence 'We investigate directly using image features from a frozen Dinov2 model as embeddings' is ambiguous; please specify that these features replace the learned patch embeddings.

Circularity Check

0 steps flagged · score 1.0 of 10

No material circularity: the text-conditioning contribution is validated by a no-prompt ablation and by external benchmarks; flagged non-circular risks are untraced split provenance for the Diff-1/Diff-2 variants and evaluation prompts hand-fitted to dataset label semantics.

full rationale

The paper's derivation chain—CSeg synthesis, joint training, then evaluation—contains no step that reduces to its own inputs by construction. The central claim is benchmarked on external datasets (PSCD [42], SYSU-CD [43], VL-CMU-CD [1]) against baselines not created by the authors, and the load-bearing role of text conditioning is demonstrated by the 'w/o Prompts' ablation (Table 7: CSeg 85.91→77.72, PSCD 52.24→10.38 IoU), which shows the prompt itself carries the disambiguating signal, so the comparison is not forced by construction. CSeg test prompts are held out ('There are 1,408 unique classes and 35,271 unique prompts not seen in train'), and training on CSeg transfers to real datasets, so the synthetic benchmark is not self-confirming. There are no self-citations: every cited prior result (Dinov2 [32], SigLip [54], DETR [5], the RSCD variants [28], PSCD, SYSU-CD, VL-CMU-CD) is external and independently published. Two caveats are correctness and fairness risks rather than demonstrated circularity. First, Section 3.6 trains on 'the Diff-1 and Diff-2 variants of PSCD and VL-CMU-CD' and Section 5 evaluates on 'the Diff-1/Diff-2 unaligned variants of PSCD and VL-CMU-CD [28] (testing robustness to viewpoint changes)' without ever stating that the variants were derived only from the original training splits; if Lin et al.'s perturbation pipeline included test-site sequences, Tables 4 and 5 would be inflated near-duplicate recognition. The paper's text neither confirms nor rules this out, so it is an unverified data-provenance assumption, not an exhibited reduction. Second, the per-dataset evaluation prompts are hand-chosen to mirror the ground-truth label vocabulary: SYSU-CD is prompted with 'urban development, suburban expansion, pre-construction groundwork, vegetation alteration, road widening, and coastal construction' (the dataset's six label classes), and VL-CMU-CD uses 'Bins, Signs, Traffic-signs, Vehicles, Refuse, Construction, Maintenance Work, Buildings' because 'these are the changes we have found to be consistent in the human labels.' This grants the text-conditioned model the label semantics at test time, but the predicted masks are in no way statistically forced by the class list, so this is a benchmarking-convenience issue, not circularity. The paper itself concedes (Sec.

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

The central claim rests on a handful of hand-tuned evaluation prompts, the unstated split provenance of the unaligned variants, and the assumption that a 0.1% label-quality sample is representative. No novel physical entities are introduced.

free parameters (4)
  • SYSU-CD evaluation prompt = "urban development, suburban expansion, pre-construction groundwork, vegetation alteration, road widening, and coastal…
    Hand-crafted to cover the dataset's label semantics; performance on SYSU-CD depends on this prompt. Listed in Section 5.1 and Fig. 6a.
  • VL-CMU-CD evaluation prompt = "Bins, Signs, Traffic-signs, Vehicles, Refuse, Construction, Maintenance Work, Buildings"
    Hand-selected by the authors as the classes consistent with human labels, due to label noise; affects VL-CMU-CD results (Section 5.1, Table 5).
  • Number of segmentation query tokens N_s = Not specified in text
    The paper does not report the number of segmentation query tokens used, despite noting they are critical (Section 3.4, Algorithm 1). This is a free architectural choice.
  • CSeg class prompt template list (45 templates) and 'all' prompt list (96 prompts) = Generated with GPT-4o and manually validated
    These templates determine the text distribution of CSeg; they are arbitrary but fixed. They influence training but are not fitted to performance.
assumptions (4)
  • domain assumption The unaligned Diff-1/Diff-2 variants from Lin et al. [28] preserve the original train/test split of PSCD and VL-CMU-CD.
    The paper trains the general model on these variants and evaluates on their test sets (Section 5.1, Tables 4/5) without stating the split provenance. If violated, results are contaminated.
  • domain assumption The CSeg label accuracy of 94.0% (based on 500 manually checked samples) is representative of the full 501k-pair dataset.
    The paper validates only ~0.1% of CSeg and propagates this accuracy to the whole dataset (Section 4.3).
  • domain assumption Frozen SigLip and DINOv2 features contain enough semantic and geometric information for text-conditioned change detection across domains.
    The architecture relies on these pretrained encoders and ablations show they are critical (Section 5.3), but the paper does not establish that the feature spaces align across the diverse domains tested.
  • domain assumption Synthetic inpainting plus affine transformation is a valid proxy for real scene changes.
    CSeg is generated by inpainting SA-1B masks with LaMa and applying affine warps (Section 4); the paper acknowledges this limitation in Section 5.4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ViewDelta: Scaling Scene Change Detection through Text-Conditioning." pith.science (2026). https://pith.science/paper/35J6JPDL

@misc{pith2026241207612,
  author       = {Pith},
  title        = {Pith review of: ViewDelta: Scaling Scene Change Detection through Text-Conditioning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/35J6JPDL}},
  note         = {Machine review of arXiv:2412.07612}
}
read the original abstract

We introduce a generalized framework for Scene Change Detection (SCD) that addresses the core ambiguity of distinguishing "relevant" from "nuisance" changes, enabling effective joint training of a single model across diverse domains and applications. Existing methods struggle to generalize due to differences in dataset labeling, where changes such as vegetation growth or lane marking alterations may be labeled as relevant in one dataset and irrelevant in another. To resolve this ambiguity, we propose ViewDelta, a text conditioned change detection framework that uses natural language prompts to define relevant changes precisely, such as a single attribute, a specific set of classes, or all observable differences. To facilitate training in this paradigm, we release the Conditional Change Segmentation dataset (CSeg), the first large-scale synthetic dataset for text conditioned SCD, consisting of over 500,000 image pairs with more than 300,000 unique textual prompts describing relevant changes. Experiments demonstrate that a single ViewDelta model trained jointly on CSeg, SYSU-CD, PSCD, VL-CMU-CD, and their unaligned variants achieves performance competitive with or superior to dataset specific models, highlighting text conditioning as a powerful approach for generalizable SCD. Our code and dataset are available at https://joshuakgao.github.io/viewdelta/.

Figures

Figures reproduced from arXiv: 2412.07612 by the authors.

Figure 1
Figure 1. Given a pair of (un)aligned images with text to con [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview diagram of the ViewDelta change detection [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Outline of the steps involved in generating CSeg. LLaVA-Next identifies object classes in an SA-1B image, which inform a text prompt and guide Grounded SAM 2 in extracting in￾stance masks. Some masks are then inpainted to simulate changes, while some original SA-1B masks are used as “red herrings”. This process yields a text prompt, Image A, Image B, and a change mask label. Training Parameters. During training, the… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Class count distributions. We separate the top 1% most frequent classes and the complete dataset for easier evalua￾tion. These figures don’t include the “all” class. Find any differences in the pumpkin, leaf. All Items. Detect all changes to the van, pedestrian. asphal…
Figure 5
Figure 5. Figure 5: Evaluation of CSeg label quality. Change mask labels are highlighted in blue. Manual inspection of the generated images show a 94% agreement with a human reviewer. 5 presents quantitative comparisons with state-of-the-art methods, qualitative analysis of model predicti…
Figure 6
Figure 6. Figure 6: Qualitative evaluation results showing model predictions across different scenarios. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 45 canonical work pages

  1. [28]

    Robust scene change detection using visual foun- dation models and cross-attention mechanisms

    Chun-Jung Lin, Sourav Garg, Tat-Jun Chin, and Feras Day- oub. Robust scene change detection using visual foun- dation models and cross-attention mechanisms. ArXiv, abs/2409.16850, 2024. 1, 2, 4, 5, 7

    arXiv 2024

  2. [39]

    The change you want to see

    Ragav Sachdeva and Andrew Zisserman. The change you want to see. 2023 IEEE/CVF Winter Conference on Applica- tions of Computer Vision (WACV), pages 3982–3991, 2022. 2, 3

    work page 2023

  3. [40]

    The change you want to see (now in 3d)

    Ragav Sachdeva and Andrew Zisserman. The change you want to see (now in 3d). In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2060– 2069, 2023. 2, 4, 5

    work page 2023

  4. [1]

    Street-view change detection with deconvolutional networks

    Pablo Fern ´andez Alcantarilla, Simon Stent, Germ ´an Ros, Roberto Arroyo, and Riccardo Gherardi. Street-view change detection with deconvolutional networks. Autonomous Robots, 42:1301 – 1322, 2016. 1, 2, 3, 5, 7

    work page 2016

  5. [2]

    Transformers in remote sens- ing: A survey

    Abdulaziz Amer Aleissaee, Amandeep Kumar, Rao Muham- mad Anwer, Salman Khan, Hisham Cholakkal, Gui-Song Xia, and Fahad Shahbaz Khan. Transformers in remote sens- ing: A survey. Remote Sensing, 15(7):1860, 2023. 2

    work page 2023

  6. [3]

    Deep learning for change detec- tion in remote sensing: a review

    Ting Bai, Le Wang, Dameng Yin, Kaimin Sun, Yepei Chen, Wenzhuo Li, and Deren Li. Deep learning for change detec- tion in remote sensing: a review. Geo-spatial Information Science, 26(3):262–288, 2023. 1

    work page 2023

  7. [4]

    A transformer-based siamese network for change detection

    Wele Gedara Chaminda Bandara and Vishal M Patel. A transformer-based siamese network for change detection. In IGARSS 2022-2022 IEEE International Geoscience and Re- mote Sensing Symposium , pages 207–210. IEEE, 2022. 4, 6

    work page 2022

  8. [5]

    End-to- end object detection with transformers, 2020

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to- end object detection with transformers, 2020. 4

    work page 2020

Show all 57 references
  1. [6]

    A spatial-temporal attention- based method and a new dataset for remote sensing image change detection

    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), 2020. 2

    2020

  2. [7]

    Remote sensing im- age change detection with transformers

    Hao Chen, Zipeng Qi, and Zhenwei Shi. Remote sensing im- age change detection with transformers. IEEE Transactions on Geoscience and Remote Sensing, 60:1–14, 2021. 6

    2021

  3. [8]

    Changemamba: Remote sensing change de- tection with spatio-temporal state space model

    Hongruixuan Chen, Jian Song, Chengxi Han, Junshi Xia, and Naoto Yokoya. Changemamba: Remote sensing change de- tection with spatio-temporal state space model. IEEE Trans- actions on Geoscience and Remote Sensing, 2024. 4, 6, 7

    2024

  4. [9]

    Dr- tanet: Dynamic receptive temporal attention network for street scene change detection

    Shuo Chen, Kailun Yang, and Rainer Stiefelhagen. Dr- tanet: Dynamic receptive temporal attention network for street scene change detection. 2021 IEEE Intelligent Vehi- cles Symposium (IV), pages 502–509, 2021. 7

    2021

  5. [10]

    When deep learning meets metric learning: Remote sensing image scene classification via learning discrimina- tive cnns

    Gong Cheng, Ceyuan Yang, Xiwen Yao, Lei Guo, and Jun- wei Han. When deep learning meets metric learning: Remote sensing image scene classification via learning discrimina- tive cnns. IEEE Transactions on Geoscience and Remote Sensing, 56:2811–2821, 2018. 4, 6

    2018

  6. [11]

    Change detection methods for remote sensing in the last decade: A comprehensive review

    Guangliang Cheng, Yunmeng Huang, Xiangtai Li, Shuchang Lyu, Zhaoyang Xu, Hongbo Zhao, Qi Zhao, and Shiming Xiang. Change detection methods for remote sensing in the last decade: A comprehensive review. Remote Sensing, 16 (13):2355, 2024. 1, 5

    2024

  7. [12]

    Re- gion filling and object removal by exemplar-based image in- painting

    Antonio Criminisi, Patrick P ´erez, and Kentaro Toyama. Re- gion filling and object removal by exemplar-based image in- painting. IEEE Transactions on image processing , 13(9): 1200–1212, 2004. 5

    2004

  8. [13]

    Fully convolutional siamese networks for change detection, 2018

    Rodrigo Caye Daudt, Bertrand Le Saux, and Alexandre Boulch. Fully convolutional siamese networks for change detection, 2018. 4, 6

    2018

  9. [14]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2010 arXiv

  10. [15]

    Foreground gating and background refining network for surveillance object detec- tion

    Zhihang Fu, Yaowu Chen, Hongwei Yong, Rongxin Jiang, Lei Zhang, and Xian-Sheng Hua. Foreground gating and background refining network for surveillance object detec- tion. IEEE Transactions on Image Processing, PP:1–1, 2019. 1

    2019

  11. [16]

    Dat- acomp: In search of the next generation of multimodal datasets

    Samir Yitzhak Gadre, Gabriel Ilharco, Alex Fang, Jonathan Hayase, Georgios Smyrnis, Thao Nguyen, Ryan Marten, Mitchell Wortsman, Dhruba Ghosh, Jieyu Zhang, et al. Dat- acomp: In search of the next generation of multimodal datasets. Advances in Neural Information Processing Sys...

    2024

  12. [17]

    A framework for the detection and attribution of biodiversity change

    Andrew Gonzalez, Jonathan M Chase, and Mary I O’Connor. A framework for the detection and attribution of biodiversity change. Philosophical Transactions of the Royal Society B , 378(1881):20220182, 2023. 2

    2023

  13. [18]

    Patel, Richard Hosfelt, Sandra Sajeev, Eric T

    Ritwik Gupta, Bryce Goodman, Nirav N. Patel, Richard Hosfelt, Sandra Sajeev, Eric T. Heim, Jigar Doshi, Keane Lu- cas, Howie Choset, and Matthew E. Gaston. xbd: A dataset for assessing building damage from satellite imagery. ArXiv, abs/1911.09296, 2019. 2

    1911 arXiv

  14. [19]

    Fast flood extent mon- itoring with sar change detection using google earth engine

    Ebrahim Hamidi, Brad G Peter, David F Mu ˜noz, Hamed Moftakhari, and Hamid Moradkhani. Fast flood extent mon- itoring with sar change detection using google earth engine. IEEE Transactions on Geoscience and Remote Sensing , 61: 1–19, 2023. 2

    2023

  15. [20]

    A long-term analysis of urbanization pro- cess, landscape change, and carbon sources and sinks: A case study in china’s yangtze river delta region

    Ji Han, Xing Meng, Xiang Zhou, Bailu Yi, Min Liu, and Wei-Ning Xiang. A long-term analysis of urbanization pro- cess, landscape change, and carbon sources and sinks: A case study in china’s yangtze river delta region. Journal of Cleaner Production, 141:1040–1050, 2017. 1

    2017

  16. [21]

    Image repairing: Robust im- age synthesis by adaptive nd tensor voting

    Jiaya Jia and Chi-Keung Tang. Image repairing: Robust im- age synthesis by adaptive nd tensor voting. In 2003 IEEE Computer Society Conference on Computer Vision and Pat- tern Recognition, 2003. Proceedings., pages I–I. IEEE, 2003. 5

    2003

  17. [22]

    A survey on deep learning-based change detection from high- resolution remote sensing images

    Huiwei Jiang, Min Peng, Yuanjun Zhong, Haofeng Xie, Zemin Hao, Jingming Lin, Xiaoli Ma, and Xiangyun Hu. A survey on deep learning-based change detection from high- resolution remote sensing images. Remote Sensing, 14(7): 1552, 2022. 1

    2022

  18. [23]

    T-rex2: Towards generic object detec- tion via text-visual prompt synergy, 2024

    Qing Jiang, Feng Li, Zhaoyang Zeng, Tianhe Ren, Shilong Liu, and Lei Zhang. T-rex2: Towards generic object detec- tion via text-visual prompt synergy, 2024. 5

    2024

  19. [24]

    Towards generalizable scene change detection

    Jaewoo Kim and Uehwan Kim. Towards generalizable scene change detection. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition, 2025. 1, 2, 7

    2025

  20. [25]

    Adam: A method for stochastic opti- mization

    Diederik P Kingma. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980, 2014. 4

    2014 arXiv

  21. [26]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In Proceedings of the IEEE/CVF international confer- ence on computer vision, pages 4015–4026, 202...

    2023

  22. [27]

    Tran- sunetcd: A hybrid transformer network for change detec- tion in optical remote-sensing images

    Qingyang Li, Ruofei Zhong, Xin Du, and Yu Du. Tran- sunetcd: A hybrid transformer network for change detec- tion in optical remote-sensing images. IEEE Transactions on Geoscience and Remote Sensing, 60:1–19, 2022. 6

    2022

  23. [29]

    Llava-next: Im- proved reasoning, ocr, and world knowledge, 2024

    Haotian Liu, Chunyuan Li, Yuheng Li, Bo Li, Yuanhan Zhang, Sheng Shen, and Yong Jae Lee. Llava-next: Im- proved reasoning, ocr, and world knowledge, 2024. 5

    2024

  24. [30]

    Grounding dino: Marrying dino with grounded pre-training for open-set object detection

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, et al. Grounding dino: Marrying dino with grounded pre-training for open-set object detection. arXiv preprint arXiv:2303.05499, 2023. 5

    2023 arXiv

  25. [31]

    Global effects of land use on local terrestrial biodiversity

    Tim Newbold, Lawrence N Hudson, Samantha LL Hill, Sara Contu, Igor Lysenko, Rebecca A Senior, Luca B ¨orger, Do- minic J Bennett, Argyrios Choimes, Ben Collen, et al. Global effects of land use on local terrestrial biodiversity. Nature, 520(7545):45–50, 2015. 1

    2015

  26. [32]

    Dinov2: Learning robust visual features with- out supervision, 2024

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mah- moud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michae...

    2024

  27. [33]

    Detecting urban changes with recurrent neural networks from multi- temporal sentinel-2 data, 2019

    Maria Papadomanolaki, Sagar Verma, Maria Vakalopoulou, Siddharth Gupta, and Konstantinos Karantzalos. Detecting urban changes with recurrent neural networks from multi- temporal sentinel-2 data, 2019. 6

    2019

  28. [34]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  29. [35]

    Zero: Memory optimizations toward training trillion parameter models

    Samyam Rajbhandari, Jeff Rasley, Olatunji Ruwase, and Yuxiong He. Zero: Memory optimizations toward training trillion parameter models. In SC20: International Confer- ence for High Performance Computing, Networking, Storage and Analysis, pages 1–16. IEEE, 2020. 4

    2020

  30. [36]

    Sam 2: Segment anything in images and videos,

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman R¨adle, Chloe Rolland, Laura Gustafson, Eric Mintun, Junt- ing Pan, Kalyan Vasudev Alwala, Nicolas Carion, Chao- Yuan Wu, Ross Girshick, Piotr Doll´ar, and Christoph Feic...

  31. [37]

    Grounding dino 1.5: Ad- vance the ”edge” of open-set object detection, 2024

    Tianhe Ren, Qing Jiang, Shilong Liu, Zhaoyang Zeng, Wen- long Liu, Han Gao, Hongjie Huang, Zhengyu Ma, Xiaoke Jiang, Yihao Chen, Yuda Xiong, Hao Zhang, Feng Li, Peijun Tang, Kent Yu, and Lei Zhang. Grounding dino 1.5: Ad- vance the ”edge” of open-set object detection, 2024

    2024

  32. [38]

    Grounded sam: Assembling open-world models for diverse visual tasks,

    Tianhe Ren, Shilong Liu, Ailing Zeng, Jing Lin, Kun- chang Li, He Cao, Jiayu Chen, Xinyu Huang, Yukang Chen, Feng Yan, Zhaoyang Zeng, Hao Zhang, Feng Li, Jie Yang, Hongyang Li, Qing Jiang, and Lei Zhang. Grounded sam: Assembling open-world models for diverse visual tasks,

  33. [41]

    Using high-resolution satellite images for post-earthquake building damage assessment: a study fol- lowing the 26 january 2001 gujarat earthquake

    Keiko Saito, Robin JS Spence, Christopher Going, and Michael Markus. Using high-resolution satellite images for post-earthquake building damage assessment: a study fol- lowing the 26 january 2001 gujarat earthquake. Earthquake spectra, 20(1):145–169, 2004. 1

    2001

  34. [42]

    Weakly supervised silhouette-based semantic scene change detec- tion, 2022

    Ken Sakurada, Mikiya Shibuya, and Weimin Wang. Weakly supervised silhouette-based semantic scene change detec- tion, 2022. 1, 2, 3, 5, 6, 7, 8

    2022

  35. [43]

    A deeply supervised attention metric-based network and an open aerial image dataset for remote sensing change detection

    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 , pages 1–16, 2021. 1, 2,...

    2021

  36. [44]

    Change detection based on artificial intel- ligence: State-of-the-art and challenges

    Wenzhong Shi, Min Zhang, Rui Zhang, Shanxiong Chen, and Zhao Zhan. Change detection based on artificial intel- ligence: State-of-the-art and challenges. Remote Sensing, 12 (10):1688, 2020. 1

    2020

  37. [45]

    Review article digital change detection techniques using remotely-sensed data

    Ashbindu Singh. Review article digital change detection techniques using remotely-sensed data. International jour- nal of remote sensing, 10(6):989–1003, 1989. 1

    1989

  38. [46]

    Automatic post-disaster damage mapping using deep-learning tech- niques for change detection: Case study of the tohoku tsunami

    J ´er´emie Sublime and Ekaterina Kalinicheva. Automatic post-disaster damage mapping using deep-learning tech- niques for change detection: Case study of the tohoku tsunami. Remote Sensing, 11(9):1123, 2019. 1

    2019

  39. [47]

    Eva-clip: Improved training techniques for clip at scale

    Quan Sun, Yuxin Fang, Ledell Wu, Xinlong Wang, and Yue Cao. Eva-clip: Improved training techniques for clip at scale. arXiv preprint arXiv:2303.15389, 2023. 3

    2023 arXiv

  40. [48]

    Resolution-robust large mask inpainting with fourier convolutions

    Roman Suvorov, Elizaveta Logacheva, Anton Mashikhin, Anastasia Remizova, Arsenii Ashukha, Aleksei Silvestrov, Naejin Kong, Harshith Goka, Kiwoong Park, and Victor Lempitsky. Resolution-robust large mask inpainting with fourier convolutions. arXiv preprint arXiv:2109.07161 ,

    arXiv

  41. [49]

    Wang, Bin-Bin Gao, and Chengjie Wang

    G. Wang, Bin-Bin Gao, and Chengjie Wang. How to re- duce change detection to semantic segmentation. ArXiv, abs/2206.07557, 2022. 2, 7

    2022 arXiv

  42. [50]

    Building damage detection using u-net with attention mechanism from pre-and post-disaster remote sensing datasets

    Chuyi Wu, Feng Zhang, Junshi Xia, Yichen Xu, Guoqing Li, Jibo Xie, Zhenhong Du, and Renyi Liu. Building damage detection using u-net with attention mechanism from pre-and post-disaster remote sensing datasets. Remote Sensing, 13 (5):905, 2021. 1

    2021

  43. [51]

    Robust image forgery detection over online social network shared 10 images

    Haiwei Wu, Jiantao Zhou, Jinyu Tian, and Jun Liu. Robust image forgery detection over online social network shared 10 images. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 13430–13439, 2022. 5

    2022

  44. [52]

    Semantic change detection with asymmetric siamese networks, 2021

    Kunping Yang, Gui-Song Xia, Zicheng Liu, Bo Du, Wen Yang, Marcello Pelillo, and Liangpei Zhang. Semantic change detection with asymmetric siamese networks, 2021. 2

    2021

  45. [53]

    Pedestrian be- havior modeling from stationary crowds with applications to intelligent surveillance

    Shuai Yi, Hongsheng Li, and Xiaogang Wang. Pedestrian be- havior modeling from stationary crowds with applications to intelligent surveillance. Trans. Img. Proc., 25(9):4354–4368,

  46. [54]

    Sigmoid loss for language image pre-training

    Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. Sigmoid loss for language image pre-training. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11975–11986, 2023. 3

    2023

  47. [55]

    Swinsunet: Pure transformer network for remote sensing im- age change detection

    Cui Zhang, Liejun Wang, Shuli Cheng, and Yongming Li. Swinsunet: Pure transformer network for remote sensing im- age change detection. IEEE Transactions on Geoscience and Remote Sensing, 60:1–13, 2022. 4, 6

    2022

  48. [56]

    Changen2: Multi-temporal re- mote sensing generative change foundation model

    Zhuo Zheng, Stefano Ermon, Dongjun Kim, Liangpei Zhang, and Yanfei Zhong. Changen2: Multi-temporal re- mote sensing generative change foundation model. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47:725–741, 2024. 2, 3

    2024

  49. [57]

    A review of multi- class change detection for satellite remote sensing imagery

    Qiqi Zhu, Xi Guo, Ziqi Li, and Deren Li. A review of multi- class change detection for satellite remote sensing imagery. Geo-spatial Information Science, 27(1):1–15, 2024. 2 11

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.