Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Semantic Localization Guiding Segment Anything Model For Reference Remote Sensing Image Segmentation

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A two-stage pipeline that converts a text description into a bounding box and lets the Segment Anything Model refine the mask outperforms end-to-end referring remote sensing segmentation models.

desk verdict A sensible two-stage framework with a genuinely useful new dataset, but the RRSIS-M SOTA claim is confounded by LQVG's DIOR-RSVG training; needs controlled comparisons before accepting the headline. read the letter →

arxiv 2506.10503 v1 pith:3NOEP6HW submitted 2025-06-12 cs.CV cs.AI

Shuyang Li , Shuang Wang , Zhuangzhuang Sun , Jing Xiao This is my paper · ORCID
classification cs.CVcs.AI
keywords referenceremotesensingimagesegmentationvisualgroundingSegmentAnythingModelprompt-guidedweaklysupervisedGrabCutboundaryrefinementRRSIS-Mdatasetclustering-basedforegroundpointgenerator
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

Reference remote sensing image segmentation asks a model to produce a segmentation mask for the object a sentence describes. The paper's central claim is that this task does not need an end-to-end multimodal segmentation network: it can be split into a coarse localization stage, where a visual grounding model converts the text into bounding-box coordinates, and a fine segmentation stage, where the Segment Anything Model (SAM) turns the box into a precise mask. This second stage can be train-free and needs no mask-level labels, so the framework's weakly supervised variant only requires bounding boxes. On the standard RRSIS-D benchmark the fully supervised version lifts mIoU from the previous state of the art's 64.22 to 70.61, and the same two-stage design also reports gains on the paper's new manually annotated RRSIS-M dataset. A sympathetic reader would take the contribution to be a demonstration that decomposing the task by semantic localization is both more accurate and much cheaper in annotation effort than single-stage fusion.

What carries the argument

The machinery is a prompt chain from text to mask. Stage one is LQVG, a DETR-style visual grounding model that fuses text and image features and outputs a bounding box around the described object. Stage two is SAM with a frozen image encoder, whose prompt encoder receives that box plus a foreground point produced by the clustering-based foreground point generator (CFPG): CFPG crops the box, runs KMeans++ to split pixels into two clusters, applies watershed segmentation to separate connected regions, and selects the region with the highest convexity as the target, then takes its geometric center as the point prompt. The SAM mask is then refined by a mask boundary iterative optimization strategy adapted from GrabCut, which builds foreground and background Gaussian mixture models and minimizes a graph-cut energy. The argument is that this chain lets SAM's boundary competence carry the fine segmentation while the grounding stage supplies the semantic localization that end-to-end networks struggle to learn from scarce remote-sensing annotations.

What would settle it

Measure LQVG's box localization accuracy (for example, box IoU against the ground-truth mask's bounding box) on the RRSIS-D and RRSIS-M test sets, and rerun PSLG-SAM with ground-truth boxes replacing the predicted ones. If the ground-truth-box version is substantially better, the claimed state-of-the-art result is carried by the grounding stage; if it is not, the segmentation stage is the limiting factor. In addition, checking the overlap between RRSIS-M test images and LQVG's DIOR-RSVG training images would settle whether the new-dataset numbers measure transfer or memorization.

Watch

Extended reading notes

Core claim

The central discovery claimed is that referring remote sensing segmentation can be reframed as localization-plus-segmentation: a remote-sensing visual grounding model (LQVG) provides prompt coordinates, and a frozen SAM, guided by a clustering-based foreground point generator and a mask boundary iterative optimization strategy, produces the final mask. The framework is reported to surpass existing state-of-the-art models on both RRSIS-D and the new RRSIS-M dataset, with the weakly supervised variant (trained on bounding boxes only) already beating the prior fully supervised RMSIN baseline in mIoU on RRSIS-D (67.41 versus 64.22). Fine-tuning only the lightweight SAM mask decoder on 10% of the mask data recovers most of the fully supervised gain (70.13 versus 70.67 mIoU in the paper's data-size experiment). A t-SNE analysis of encoder outputs is offered as evidence that SAM's frozen features separate remote-sensing categories more cleanly than the multi-scale features of the previous best method.

Load-bearing premise

The framework assumes the pretrained LQVG visual grounding model already knows how to find the text-described object and frame it with a usable box in remote sensing images; the paper never measures LQVG's grounding accuracy on RRSIS-D or RRSIS-M, and if those boxes drift, the clustering point generator and SAM cannot recover.

Editorial extensions

If this is right

  • Referring remote sensing segmentation can be performed with bounding-box-only supervision; the weakly supervised variant already exceeds the prior fully supervised baseline in mIoU on RRSIS-D (67.41 versus 64.22).
  • The segmentation stage can be kept frozen, so adapting the framework to a new region or sensor costs only the grounding model, not a mask-training run.
  • Fine-tuning only the lightweight SAM mask decoder on 10% of the mask data recovers most of the gain of full supervision (70.13 versus 70.67 mIoU at 100% in the paper's data-size experiment), so large mask-annotation campaigns have diminishing returns once localization is solved.
  • Because the stages are independent, improving the visual grounding model or the prompt-generation modules should directly improve final masks without re-designing the segmentation pipeline.

Reading between the lines

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

  • An experiment the paper does not run would apportion credit: replace LQVG boxes with ground-truth boxes of matched tightness; a large mIoU jump would show the reported gains are mainly the grounding stage's contribution rather than SAM's or the prompt modules'.
  • The paper's RRSIS-M test set is built from DIOR-RSVG imagery and descriptions, while LQVG is trained on DIOR-RSVG; checking whether test images overlap the grounding training split would clarify whether the RRSIS-M results reflect generalization or near-duplicate transfer.
  • The convexity heuristic inside CFPG assumes the target is the most convex connected region in the box, so thin, elongated, or partially occluded objects are a plausible failure mode worth a targeted stress test.
  • If the two-stage recipe transfers beyond the specific models used here (for example, different grounding backbones or different SAM checkpoints), it would suggest that referring segmentation in general domains can be decomposed into localization and boundary refinement rather than learned end-to-end.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The manuscript proposes PSLG-SAM, a two-stage framework for reference remote sensing image segmentation (RRSIS). In the first stage, a visual grounding model (LQVG) converts an image-text pair into a bounding-box coordinate prompt. In the second stage, the bounding box crops the region of interest, a clustering-based foreground point generator (CFPG) supplies a center point, and the Segment Anything Model (SAM) with a frozen image encoder and prompt encoder produces a mask; a mask boundary iterative optimization (MBO) strategy, inspired by GrabCut, refines the mask edges. The mask decoder can be fine-tuned or kept frozen, yielding fully supervised and weakly supervised variants. The paper also introduces RRSIS-M, a manually annotated benchmark derived from the DIOR-RSVG dataset, and reports experiments on RRSIS-D and RRSIS-M that claim state-of-the-art performance in both settings.

Significance. The two-stage decomposition is conceptually attractive and practically relevant: it leverages strong pretrained components (visual grounding and SAM) and offers a train-free segmentation stage that reduces mask-annotation demand. The reported RRSIS-D fully supervised mIoU improvement (70.61 vs. 64.22 for RMSIN) is substantial, and the RRSIS-M dataset, if carefully constructed, could be a useful contribution to the community. The paper also provides ablation evidence that CFPG and MBO contribute to the final scores, and it promises public code. However, the central claim of state-of-the-art performance is currently conditional on an unverified data-separation assumption: RRSIS-M is built from DIOR-RSVG, and LQVG is a DIOR-RSVG-era visual grounding model, so the evaluation may involve train/test overlap in the grounding stage. Until this is resolved, the significance of the empirical results is not fully established.

major comments (4)
  1. [§2.2 and §4.1] The RRSIS-M evaluation has a potential train/test overlap that is not addressed. Section 4.1 states that RRSIS-M was constructed by "leveraging bounding box prompts and textual descriptions derived from the DIOR-RSVG dataset," while Section 2.2 says the framework "adopts LQVG" to generate coordinate prompts, and LQVG is a visual grounding model developed on DIOR-RSVG. The paper does not report the DIOR-RSVG split used to train LQVG, nor does it state whether the RRSIS-M test images and their text-box pairs were excluded from LQVG's training data. If they were not excluded, the bounding boxes fed to CFPG and SAM in Tables 2 and 4 are not produced by genuine grounding transfer, and the weakly supervised results—where the LQVG box is the only test-time localization signal—would be invalid as evidence for the method. The same concern applies to RRSIS-D if its source imagery overlaps with DIOR-RSVG, which the paper also does not discuss.
  2. [§4.4, Tables 1 and 2] The comparison with state-of-the-art methods is not confound-controlled. PSLG-SAM uses a pretrained SAM and a pretrained LQVG, while the baseline methods (RRN, CMSA, LAVT, RMSIN, etc.) are trained from scratch on the task-specific mask labels. The reported SOTA claim therefore conflates the contribution of the proposed framework with the contribution of additional pretrained grounding knowledge and SAM's large-scale pretraining. A controlled experiment—for example, feeding LQVG's bounding boxes as external prompts to a baseline segmentation model, or fine-tuning all methods with the same pretrained backbone—is needed to attribute the performance gain to the proposed components.
  3. [§4.4, Table 1] The weakly supervised claim is only partially supported on RRSIS-D. In Table 1, PSLG-SAM* achieves mIoU 67.41 versus 64.22 for RMSIN, but its oIoU is 75.40 versus 77.53, i.e., 2.1 points lower. The text says the method achieved "a 3.2% improvement in mIoU and a 2.1% difference in oIoU" and later acknowledges the lower oIoU, but the abstract and conclusion present the weakly supervised variant as a success without this caveat. The authors should state plainly that the weakly supervised model improves mIoU but not oIoU on RRSIS-D, and justify why mIoU should be treated as the primary metric for the SOTA claim.
  4. [§3.2, Eq. (8)] The CFPG module relies on the assumption that the target foreground object is the connected component with the greatest convexity within the bounding box. This is stated as an observation that "almost all" foreground objects have smaller convexity defects than background regions, but no quantitative validation is provided. For concave or complex targets such as an expressway service area or scattered buildings, the largest-convexity selection could pick a background region. A failure analysis or an experiment with oracle bounding boxes would clarify how often this assumption holds and how much of CFPG's reported gain depends on it.
minor comments (6)
  1. [Abstract and Introduction] There are several typos, including "aassist" in the contribution list and "RRISS" in Figure 1; these should be corrected.
  2. [§3.3] The mask boundary optimization is described as an energy minimization with a data term and a smoothness term, but the smoothness term is never written as an equation and the weighting between the terms is not specified. The authors should provide the full energy function and the parameter values for reproducibility.
  3. [§4.2] It is unclear whether LQVG is trained from scratch or fine-tuned from a released checkpoint, and on which data split. This is directly relevant to the data-leakage concern and should be stated explicitly.
  4. [Table 1 and Table 2] The column header "Pr@0.5" appears as "P r@0.5" and the metric is sometimes written "MIoU" and sometimes "mIoU"; the caption should define all abbreviations consistently.
  5. [§4.4] The sentence describing the weakly supervised RRSIS-D result as a "2.1% difference in oIoU" is confusing because the difference is a decrease, not an improvement; please rephrase to state the direction explicitly.
  6. [§5] The conclusion claims the framework is a "scalable, efficient solution" and "offering promising potential," but it does not mention the reliance on LQVG's pretrained knowledge or the possibility that LQVG boxes can be inaccurate for unusual remote sensing targets. A limitation paragraph would improve the paper's balance.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: PSLG-SAM's segmentation outputs are evaluated against manually annotated masks, and the visual-grounding boxes are intermediate prompts, not fitted predictions renamed as results.

full rationale

The paper's claimed derivation chain—visual grounding boxes feeding SAM prompts, with CFPG and MBO refining masks—does not reduce to its inputs by construction. The final mask outputs are produced by SAM's mask decoder from image and prompt features (Eq. 3) and compared against manually annotated masks in RRSIS-D and RRSIS-M; no fitted parameter from the segmentation stage is renamed as a prediction. LQVG is presented as an existing external grounding approach (Section 2.2), not a self-citation, and the bounding-box prompts are intermediate cues rather than the evaluation target. The strongest potential concern is dataset-source overlap: RRSIS-M was constructed from DIOR-RSVG (Section 4.1) and the grounding model is likely trained on DIOR-RSVG, so RRSIS-M test boxes could be in-distribution; however, the paper does not report the DIOR-RSVG train/test split used for the grounding model, so overlap cannot be excluded. That is a benchmark-contamination and evaluation-fairness risk, especially for the weakly supervised setting, not a circularity of the derivation chain: the mask labels themselves were manually produced and are not used to train the grounding model or the frozen SAM encoder. No enumerated circularity step is therefore established, and the central benchmark results retain independent content.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The ledger shows the method's dependence on hand-chosen hyperparameters (KMeans k, area threshold, watershed threshold, GMM components) and on two domain-level heuristics: the convexity rule and LQVG's transferability. No new physical entities are introduced.

free parameters (4)
  • KMeans cluster count k = 2
    The number of clusters in CFPG is fixed to 2 (foreground/background) in Section 3.2. The central point prompt depends on this choice.
  • area threshold for connected region selection = not specified
    Equation (8) filters regions by Area(Ri) > Threshold, but the threshold value is never stated. The foreground point location depends on it.
  • watershed distance-transform threshold = not specified
    Section 3.2 states 'Foreground markers are generated by thresholding the distance transform with a threshold' but the threshold is not given.
  • GMM component count K in MBO = not specified
    Equation (10) uses K Gaussian components for foreground/background; K is not reported. This affects boundary optimization outcomes.
assumptions (3)
  • ad hoc to paper The connected region with the greatest convexity within the ROI is the target foreground object, used in Eq. (8) to select Rbest.
    Section 3.2 states 'we observed that almost all foreground objects have smaller convexity defects than the background regions' without quantitative evidence. If this fails, the foreground point prompt is wrong.
  • domain assumption LQVG, trained on DIOR-RSVG, produces bounding boxes accurate enough on RRSIS-D and RRSIS-M to serve as ROI prompts.
    The framework relies on the grounding coordinates from Eq. (1) as the only localization signal; no fine-tuning of LQVG on RRSIS data is described.
  • standard math GrabCut-style energy minimization with GMMs converges to the correct boundary in remote sensing scenes.
    Section 3.3 applies the classical GrabCut iterative graph cut. This is an established algorithm, but its convergence and correctness on aerial imagery is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semantic Localization Guiding Segment Anything Model For Reference Remote Sensing Image Segmentation." pith.science (2026). https://pith.science/paper/3NOEP6HW

@misc{pith2026250610503,
  author       = {Pith},
  title        = {Pith review of: Semantic Localization Guiding Segment Anything Model For Reference Remote Sensing Image Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3NOEP6HW}},
  note         = {Machine review of arXiv:2506.10503}
}
read the original abstract

The Reference Remote Sensing Image Segmentation (RRSIS) task generates segmentation masks for specified objects in images based on textual descriptions, which has attracted widespread attention and research interest. Current RRSIS methods rely on multi-modal fusion backbones and semantic segmentation heads but face challenges like dense annotation requirements and complex scene interpretation. To address these issues, we propose a framework named \textit{prompt-generated semantic localization guiding Segment Anything Model}(PSLG-SAM), which decomposes the RRSIS task into two stages: coarse localization and fine segmentation. In coarse localization stage, a visual grounding network roughly locates the text-described object. In fine segmentation stage, the coordinates from the first stage guide the Segment Anything Model (SAM), enhanced by a clustering-based foreground point generator and a mask boundary iterative optimization strategy for precise segmentation. Notably, the second stage can be train-free, significantly reducing the annotation data burden for the RRSIS task. Additionally, decomposing the RRSIS task into two stages allows for focusing on specific region segmentation, avoiding interference from complex scenes.We further contribute a high-quality, multi-category manually annotated dataset. Experimental validation on two datasets (RRSIS-D and RRSIS-M) demonstrates that PSLG-SAM achieves significant performance improvements and surpasses existing state-of-the-art models.Our code will be made publicly available.

Figures

Figures reproduced from arXiv: 2506.10503 by the authors.

Figure 1
Figure 1. In comparison to the proposed PSLG-SAM with tradi [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The pipeline of proposed PSLG-SAM. A visual grounding model is introduced to transform the reference image segmentation into [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. To facilitate integration with prevailing referring [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 3
Figure 3. Figure 3: Qualitative comparison of mask annotation quality between RRSIS-D (a) and RRSIS-M (b). [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]
Figure 4
Figure 4. Figure 4: Qualitative comparisons between proposed method and the previous methods RMSIN and LAVT. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: t-SNE visualization of Encoder output features between proposed method and the previous methods RMSIN. The output features [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. DiCoR: Decoupled Referent Disambiguation and Contour Recalibration for Efficient Referring Remote Sensing Image Segmentation

    cs.CV 2026-08 conditional novelty 6.0 of 10

    DiCoR combines candidate-ranking referent disambiguation with residual contour recalibration on an efficient joint fusion backbone, setting new state-of-the-art scores on RefSegRS, RRSIS-D, and RISBench.

Reference graph

Works this paper leans on

30 extracted references · 29 canonical work pages · cited by 1 Pith paper

  1. [1]

    [Arthur and Vassilvitskii(2007)] David Arthur and Sergei Vassilvitskii

    work page 2007

  2. [3]

    [Carion et al.(2020)] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko

    work page 2020

  3. [8]

    In Proceedings of the IEEE conference on computer vision and pattern recognition

    Natural language object retrieval. In Proceedings of the IEEE conference on computer vision and pattern recognition. 4555–4564. [Hu et al.(2020)] Zhiwei Hu, Guang Feng, Jiayu Sun, Lihe Zhang, and Huchuan Lu

    work page 2020

  4. [9]

    In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Bi-directional relation- ship inferring network for referring image segmenta- tion. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 4424–4433. [Huang et al.(2020)] Shaofei Huang, Tianrui Hui, Si Liu, Guanbin Li, Yunchao Wei, Jizhong Han, Luoqi Liu, and Bo Li

    work page 2020

  5. [10]

    In Proceedings of the IEEE/CVF conference on computer vision and pat- tern recognition

    Referring image segmentation via cross- modal progressive comprehension. In Proceedings of the IEEE/CVF conference on computer vision and pat- tern recognition. 10488–10497. [Hui et al.(2020)] Tianrui Hui, Si Liu, Shaofei Huang, Guan- bin Li, Sansi Yu, Faxi Zhang, and Jizhong Han

    work page 2020

  6. [11]

    In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part X 16

    Linguistic structure guided context modeling for refer- ring image segmentation. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part X 16 . Springer, 59–75. [Klemas(2013)] Victor Klemas

    work page 2013

  7. [13]

    In Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition

    Referring image segmentation via recurrent refinement networks. In Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition . 5745–5753. [Liu et al.(2017)] Chenxi Liu, Zhe Lin, Xiaohui Shen, Jimei Yang, Xin Lu, and Alan Yuille

    work page 2017

  8. [15]

    IEEE Transactions on Pattern Analysis and Machine In- telligence 44, 9 (2021), 4761–4775

    Cross-modal progressive comprehension for referring segmentation. IEEE Transactions on Pattern Analysis and Machine In- telligence 44, 9 (2021), 4761–4775. [Liu et al.(2024)] Sihan Liu, Yiwei Ma, Xiaoqing Zhang, Haowei Wang, Jiayi Ji, Xiaoshuai Sun, and Rongrong Ji

    work page 2021

Show all 30 references
  1. [18]

    In Proceedings of the European Conference on Computer Vision (ECCV)

    Dy- namic multimodal instance segmentation guided by nat- ural language queries. In Proceedings of the European Conference on Computer Vision (ECCV). 630–645. [Meyer(1992)] F. Meyer

    1992

  2. [23]

    [Yang et al.(2013)] Jun Yang, Peng Gong, Rong Fu, Minghua Zhang, Jingming Chen, Shunlin Liang, Bing Xu, Jiancheng Shi, and Robert Dickinson

    Remote sensing in urban planning: Contributions towards eco- logically sound policies? Landscape and urban plan- ning 204 (2020), 103921. [Yang et al.(2013)] Jun Yang, Peng Gong, Rong Fu, Minghua Zhang, Jingming Chen, Shunlin Liang, Bing Xu, Jiancheng Shi, and Robert Dickinson

    2020

  3. [24]

    Nature climate change 3, 10 (2013), 875–883

    The role of satellite remote sensing in climate change stud- ies. Nature climate change 3, 10 (2013), 875–883. [Yang et al.(2022)] Zhao Yang, Jiaqi Wang, Yansong Tang, Kai Chen, Hengshuang Zhao, and Philip HS Torr

    2013

  4. [25]

    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion

    Lavt: Language-aware vision transformer for referring image segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion. 18155–18165. [Ye et al.(2022)] Jiabo Ye, Junfeng Tian, Ming Yan, Xi- aoshan Yang, Xuwu Wang, Ji Zhang, Liang He, an...

    2022

  5. [26]

    InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition

    Shifting more attention to visual backbone: Query-modulated refinement networks for end-to-end visual grounding. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition . 15502–15512. [Ye et al.(2019)] Linwei Ye, Mrigank Rochan, Zhi Liu, and Yang Wang

    2019

  6. [28]

    In Proceedings of the IEEE conference on computer vision and pattern recognition

    Mattnet: Modular attention network for referring ex- pression comprehension. In Proceedings of the IEEE conference on computer vision and pattern recognition . IEEE, 1307–1315. [Yuan et al.(2024)] Zhenghang Yuan, Lichao Mou, Yuan- sheng Hua, and Xiao Xiang Zhu

    2024

  7. [29]

    IEEE Transactions on Geoscience and Remote Sensing (2024)

    Rrsis: Referring remote sensing image segmentation. IEEE Transactions on Geoscience and Remote Sensing (2024). [Zhan et al.(2023)] Yang Zhan, Zhitong Xiong, and Yuan Yuan

    2024

  8. [30]

    IEEE Transactions on Geoscience and Remote Sensing 61 (2023), 1–13

    Rsvg: Exploring data and models for visual grounding on remote sensing data. IEEE Transactions on Geoscience and Remote Sensing 61 (2023), 1–13

    2023

  9. [1992]

    In 1992 International Conference on Image Processing and its Applications

    Color image segmentation. In 1992 International Conference on Image Processing and its Applications. 303–306. [Pham et al.(2011)] Hai Minh Pham, Yasushi Yamaguchi, and Thanh Quang Bui

    2011

  10. [2004]

    ACM trans- actions on graphics (TOG) 23, 3 (2004), 309–314

    ” GrabCut” interactive fore- ground extraction using iterated graph cuts. ACM trans- actions on graphics (TOG) 23, 3 (2004), 309–314. [Sun et al.(2022)] Yuxi Sun, Shanshan Feng, Xutao Li, Yun- ming Ye, Jian Kang, and Xu Huang

    2004

  11. [2007]

    In Proceedings of the Eighteenth An- nual ACM-SIAM Symposium on Discrete Algorithms, SODA 2007, New Orleans, Louisiana, USA, January 7- 9,

    K-Means++: The Advantages of Careful Seeding. In Proceedings of the Eighteenth An- nual ACM-SIAM Symposium on Discrete Algorithms, SODA 2007, New Orleans, Louisiana, USA, January 7- 9,

    2007

  12. [2011]

    Landscape and Urban Planning 100, 3 (2011), 223–230

    A case study on the re- lation between city planning and urban growth using re- mote sensing and spatial metrics. Landscape and Urban Planning 100, 3 (2011), 223–230. [Rother et al.(2004)] Carsten Rother, Vladimir Kolmogorov, and Andrew Blake

    2011

  13. [2013]

    Fisheries Research 148 (2013), 124–136

    Fisheries applications of remote sensing: An overview. Fisheries Research 148 (2013), 124–136. [Li et al.(2018)] Ruiyu Li, Kaican Li, Yi-Chun Kuo, Michelle Shu, Xiaojuan Qi, Xiaoyong Shen, and Jiaya Jia

    2013

  14. [2016]

    In Proceedings of the IEEE conference on computer vision and pattern recognition

    Deep residual learning for im- age recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition . 770–778. [Hu et al.(2016)] Ronghang Hu, Huazhe Xu, Marcus Rohrbach, Jiashi Feng, Kate Saenko, and Trevor Darrell

    2016

  15. [2017]

    In Proceedings of the IEEE international conference on computer vision

    Recurrent multi- modal interaction for referring image segmentation. In Proceedings of the IEEE international conference on computer vision. 1271–1280. [Liu et al.(2021)] Si Liu, Tianrui Hui, Shaofei Huang, Yun- chao Wei, Bo Li, and Guanbin Li

    2021

  16. [2018]

    arXiv preprint arXiv:1810.04805 (2018)

    Bert: Pre-training of deep bidirectional transformers for language under- standing. arXiv preprint arXiv:1810.04805 (2018). [He et al.(2016)] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun

    2018 arXiv

  17. [2019]

    InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition

    Cross-Modal Self-Attention Net- work for Referring Image Segmentation. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 10502–10511. [Yu et al.(2018)] Licheng Yu, Zhe Lin, Xiaohui Shen, Jimei Yang, Xin Lu, Mohit Bansal, and Tamara L Berg

    2018

  18. [2020]

    In European conference on computer vision

    End-to-end object detection with transformers. In European conference on computer vision. Springer, 213–229. [Deng et al.(2021)] Jiajun Deng, Zhengyuan Yang, Tianlang Chen, Wengang Zhou, and Houqiang Li

    2021

  19. [2021]

    In Pro- ceedings of the IEEE/CVF International Conference on Computer Vision

    Transvg: End-to-end visual grounding with transformers. In Pro- ceedings of the IEEE/CVF International Conference on Computer Vision. 1769–1779. [Devlin(2018)] Jacob Devlin

    2018

  20. [2022]

    In Proceedings of the 30th ACM International Conference on Multimedia

    Visual grounding in remote sensing images. In Proceedings of the 30th ACM International Conference on Multimedia . 404–412. [Wellmann et al.(2020)] Thilo Wellmann, Angela Lausch, Erik Andersson, Sonja Knapp, Chiara Cortinovis, Jes- sica Jache, Sebastian Scheuer, Peleg Kremer, ...

    2020

  21. [2023]

    In Proceedings of the 31st ACM International Confer- ence on Multimedia

    CARIS: Context-aware referring image segmentation. In Proceedings of the 31st ACM International Confer- ence on Multimedia. 779–788. [Margffoy-Tuay et al.(2018)] Edgar Margffoy-Tuay, Juan C P´erez, Emilio Botero, and Pablo Arbel ´aez

    2018

  22. [2024]

    InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Rotated multi-scale interaction network for re- ferring remote sensing image segmentation. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 26658–26668. [Liu et al.(2023)] Sun-Ao Liu, Yiheng Zhang, Zhaofan Qiu, Hongtao Xie, Yongdong Zhang,...

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.