Pith. sign in

REVIEW 4 major objections 7 minor 1 cited by

RemoteSAM: Towards Segment Anything for Earth Observation

T0 review · 4 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read RemoteSAM proposes that a single referring-expression-segmentation model, trained on 270K remote-sensing triplets, can unify pixel-, region-, and image-level perception tasks and beat billion-parameter vision-language models on several…

desk verdict Genuinely interesting paradigm, but the benchmark numbers are compromised by train/test overlap and an internal inconsistency. read the letter →

arxiv 2505.18022 v3 pith:6JCUHFYM submitted 2025-05-23 cs.CV

Liang Yao , Fan Liu , Delong Chen , Chuanyi Zhang , Yijun Wang , Ziyun Chen , Wei Xu , Shimin Di
show 1 more author
Yuhui Zheng
This is my paper
classification cs.CV
keywords referringexpressionsegmentationEarthobservationfoundationmodelremotesensingtaskunificationimage-text-maskdatasetmulti-labelclassificationvisualgroundingzero-shot
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 argues that referring expression segmentation—asking a model to paint a pixel mask for whatever a text prompt describes—can serve as the single primitive for all vision-centric Earth observation tasks. RemoteSAM therefore replaces the usual menu of task-specific heads or large language models with one fusion-decoder that outputs masks and converts them into boxes, labels, counts, and captions. To train it, the authors built RemoteSAM-270K, the largest remote-sensing referring-segmentation dataset to date, with 270K image-text-mask triplets covering 297 categories and 16 attribute types, generated by a scalable automatic pipeline using large vision-language models and pseudo-label filtering. If the results hold, a single 180M-parameter architecture can match or beat billion-parameter specialist systems on segmentation, grounding, and multi-label classification while remaining much cheaper to run.

What carries the argument

The load-bearing device is the referring-expression-segmentation interface: input an image and a free-form text prompt, output a binary mask. The paper defines a generalized triplet regime—one-to-one, one-to-many (all instances of a category), and one-to-zero (absent categories, to suppress spurious masks)—which lets one model answer "segment this thing," "find everything of this class," and "this class is not here." Around that mask the authors wrap deterministic converters: mask-to-bbox for grounding, contour refinement plus mask-to-bbox for object detection, per-class mask aggregation for semantic segmentation, and max-plus-average confidence pooling for classification and counting.

What would settle it

Train RemoteSAM-270K after deleting every image that overlaps the RRSISD, RisBench, DOTA, DIOR, and iSAID test splits, then rerun the reported benchmarks; if the mIoU gains (71.75 and 71.46) shrink to parity with prior methods, the SOTA claim is explained by train/test leakage. A public check could also compare image hashes between the released dataset and the official test sets.

Watch

Extended reading notes

Core claim

RemoteSAM's central claim is that pixel-level masks are the fundamental output unit of vision, so every region- and image-level task can be derived from a mask by deterministic post-processing. The model is trained only for referring expression segmentation on RemoteSAM-270K, and at inference it turns the mask into a bounding box (mask-to-bbox), a semantic map (aggregating per-class masks), a multi-label prediction (pooling class confidence), an object count, and a caption. On benchmarks, it reports 71.75 mIoU on RRSISD and 71.46 on RisBench, surpassing prior referring-segmentation methods by more than 3 points; it also reports 94% and 95% multi-label accuracy on DIOR and DOTAv2, 74.36 AP50 on RSVG grounding, 62.0% counting accuracy on DIOR, and state-of-the-art zero-shot semantic segmentation on Potsdam (91.80 mF1), all with 180M parameters.

Load-bearing premise

The central numbers assume the RRSISD and RisBench test images never appear in RemoteSAM-270K's training triplets, since the paper builds the dataset from those same benchmarks and reports no held-out split.

Editorial extensions

If this is right

  • If RemoteSAM is right, a single foundation model can serve pixel-, region-, and image-level remote sensing tasks without per-task decoders or LLM backbones.
  • The scale and semantic breadth of RemoteSAM-270K is claimed to be the main driver of generalization: more categories and attributes transfer to unseen datasets, as shown by zero-shot classification on SATIN and open-vocabulary segmentation on Vaihingen, UDD5, and DeepGlobe.
  • Because mask-based outputs are deterministic, the model's behavior on detection, counting, and classification is directly traceable to its segmentation quality; errors in masks become errors in all downstream tasks.
  • The parameter efficiency claim becomes concrete for deployment: 180M parameters versus 7B for vision-language rivals, which is material when processing high-resolution remote sensing imagery.
  • The RSVocab-1K vocabulary provides a reusable instrument for measuring whether any remote sensing dataset has enough category and attribute coverage to support foundation-model training.

Reading between the lines

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

  • Left implicit: the mask-first interface could extend to multispectral, multitemporal, or radar imagery without changing the architecture, since band-specific or time-specific attributes can be expressed in the text prompt.
  • Testable extension: applying the same mask-conversion pipeline to natural-image referring-segmentation benchmarks would reveal whether task unification is a general vision principle or specific to remote sensing.
  • An unstated corollary: because image-level outputs are pooled from masks, referring-segmentation mIoU is the single health metric; gains there should propagate upward to detection, counting, and classification, making segmentation accuracy a sufficient optimization target.
  • The paper's transfer claims would be strengthened by explicitly reporting which images from RRSISD, RisBench, DOTA, DIOR, and iSAID were excluded from RemoteSAM-270K, since the dataset is built directly from those benchmarks.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The paper proposes RemoteSAM, a referring-expression-segmentation (RES) foundation model for Earth observation, together with a new large-scale dataset RemoteSAM-270K built by integrating existing remote sensing segmentation and detection annotations and by using VLM-generated expressions. The authors claim that a single 180M-parameter model, without task-specific heads, can unify pixel-level (referring and semantic segmentation), region-level (grounding, detection), and image-level (multi-label classification, counting, captioning) tasks, and they report state-of-the-art results on RRSISD, RisBench, iSAID, DIOR, DOTAv2, and Potsdam.

Significance. If the empirical claims were valid, the work would be significant: the RES-centered unification paradigm is conceptually clean, the dataset is substantially larger than prior referring segmentation resources, and the reported parameter efficiency (180M versus 7B VLMs) is attractive. The authors also ship models and data publicly, and they provide a hierarchical remote sensing vocabulary (RSVocab-1K) as a coverage-analysis tool. These are real strengths. However, the central empirical evidence is compromised by apparent train/test overlap between RemoteSAM-270K and the benchmarks used for the headline tables, and by hyperparameter selection on the same evaluation sets. The claimed SOTA results are therefore not currently supported.

major comments (4)
  1. [Section 3.1.1 and Tables 3, 4, 6, 7] The evaluation is not held-out with respect to the training corpus. Section 3.1.1 states that RemoteSAM-270K is built by directly integrating existing referring segmentation annotations from RefSegRS, RRSISD, and RisBench, and by converting detection/instance masks from DOTA, iSAID, DIOR, LoveDA, and Potsdam into mask triplets. Tables 3, 4, 6, and 7 then report performance on RRSISD, RisBench, iSAID, Potsdam, DIOR, and DOTAv2 without disclosing any train/test split. The headline numbers (e.g., 71.75 mIoU on RRSISD, 71.46 mIoU on RisBench, 64.72 mIoU on iSAID, 94%/95% multi-label accuracy on DIOR/DOTAv2) are indistinguishable from recognition of training images. The paper must provide per-image split provenance and re-run the evaluation under a protocol where no benchmark image appears in RemoteSAM-270K.
  2. [Appendix C, Table 14 vs Table 3] There is an internal inconsistency in the reported performance of the same architecture on the same dataset. Table 14 states that BERT+Swin-Base achieves 76.21 oIoU / 64.79 mIoU on RRSISD, while Table 3 reports RemoteSAM, which uses BERT+Swin-Base per Section 5.1.1, at 80.04 oIoU / 71.75 mIoU on RRSISD. The manuscript does not explain the protocol difference (e.g., training on a subset versus the full RemoteSAM-270K, different epochs, or different evaluation splits). This discrepancy undermines the reproducibility of the main table and must be resolved before the results can be trusted.
  3. [Section 4.3, Eq. (3), Fig. 22, Table 6] Hyperparameters are tuned on evaluation data. The balance factor lambda in Eq. (3) is selected by experiments on DIOR and UCM shown in Fig. 22, and the resulting model is then evaluated on DIOR and DOTAv2 in Table 6. Likewise, the classification threshold tau_cls in Eq. (4) is set on the same benchmarks. This means the reported multi-label and image classification accuracies partly reflect thresholds fitted to the test set, inflating the measured performance. The authors should either fix these hyperparameters a priori or tune them on a separate validation split, and then report results on a truly disjoint test set.
  4. [Appendix C, Table 16 vs Table 6] A large discrepancy between two tables reporting the same task is unexplained. Table 16 reports that the prob-level multi-label classification strategy achieves 94.042% on DIOR and 75.752% on DOTAv2, whereas Table 6 reports RemoteSAM at 94% and 95% on the same datasets. If these numbers came from different protocols (e.g., different label sets, different image subsets, or different thresholds), the paper must say so; otherwise the inconsistency calls both tables into question.
minor comments (7)
  1. [Section 3.2] The vocabulary is named 'RSVocab-1K' in the text but appears as 'RSBocab-1K' in one instance; please fix the typo.
  2. [Section 3.1.2] The word 'Supplymentary' should be 'Supplementary'.
  3. [Appendix C] The phrase 'lass-wise probability aggregation function' should be 'class-wise probability aggregation function'.
  4. [Fig. 22] The right panel of Fig. 22 is labeled 'UCM' but the text does not state which task or metric it reports (e.g., multi-label accuracy or scene classification accuracy); please clarify the axis and the task.
  5. [Section 5.1.2] The evaluation setup lists DOTA, DIOR, iSAID, Potsdam, and RSVG but omits RRSISD and RisBench, which are used in Table 3 and the appendix; the dataset list should be complete.
  6. [Table 1] The column '#Attr/Sample' for RemoteSAM-270K is 3.17, but the paper also reports 16 attribute types; please define whether this is the average number of attributes per expression and clarify the counting method.
  7. [Table 4] The 'Pre-trained Data' column for RemoteSAM is 'RemoteSAM-270K', which includes iSAID and Potsdam images; since these are also the evaluation datasets in the same table, a footnote is needed to state that these results are not held-out.

Circularity Check

2 steps flagged · score 6.0 of 10

Benchmark predictions are partly fitted inputs: RRSISD/RisBench annotations are integrated into RemoteSAM-270K and then evaluated with no disclosed held-out split, and DIOR accuracy uses lambda tuned on DIOR.

  1. fitted input called prediction [Section 3.1.1 (One-to-One strategy) vs Section 5.1.2 and Table 3]
    "'One-to-One': Directly integrate existing referring segmentation annotations from RefSegRS [79], RRSISD [43], and RisBench [17]. ... We compared our approach against previous SOTA referring segmentation models across two large-scale datasets: RRSISD [43] and RisBench [17]."

    The training set RemoteSAM-270K is explicitly built by integrating the RRSISD and RisBench ground-truth referring annotations as training triplets. The same two datasets are then used as the evaluation benchmarks in Table 3, with no disclosed train/test split. The reported SOTA values (RRSISD 71.75 mIoU, RisBench 71.46 mIoU) are therefore measured on data whose annotations entered the training set: the 'prediction' is recognition of training samples rather than held-out generalization. This is further aggravated by the internal inconsistency that Table 14 reports the same BERT+Swin-Base architecture at 76.21 oIoU / 64.79 mIoU on RRSISD while Table 3 gives 80.04 oIoU / 71.75 mIoU for the same architecture and dataset, indicating an undisclosed protocol difference.

  2. fitted input called prediction [Appendix C, Fig. 22 and Eq. 3 vs Table 6]
    "To ascertain the optimal balance factor values for classification, we conduct experiments by varying the balance factor λ within the lass-wise probability aggregation function (Eq.3). The experimental results are illustrated in Fig. 22. The results indicate that when λ is set to 0.5 and 1, Multi-Label Classification and Image Classification achieved the highest accuracy, respectively. Therefore, we adopt this set of parameters for our final test results."

    The pooling balance parameter λ in Eq. 3 is selected by optimizing accuracy on DIOR (shown in Fig. 22, where the DIOR accuracy curve is plotted against λ) and on UCM. Table 6 then reports RemoteSAM's 94% multi-label accuracy on DIOR as a benchmark result. Because the free parameter governing the classification decision was fitted to the evaluation set, the reported DIOR number is a fitted value rather than a genuine prediction; the same issue applies to the threshold τ_cls defaulting to 0.5 in Eq. 4, which was not validated on a held-out split.

full rationale

The central SOTA claim is partly circular by construction. Section 3.1.1 states that the RemoteSAM-270K training set is built by directly integrating existing referring segmentation annotations from RefSegRS, RRSISD, and RisBench, yet Section 5.1.2 and Table 3 evaluate on RRSISD and RisBench without disclosing a held-out split. Consequently, the headline referring-segmentation results are not independent evidence of generalization. Similarly, the multi-label DIOR accuracy in Table 6 depends on λ tuned on DIOR in Fig. 22, so that number is a fitted value. These issues reduce the force of the paper's 'new SoTA' claim on those specific benchmarks. However, not all evidence is circular: Table 2 evaluates semantic segmentation on unseen datasets (Vaihingen, UDD5, DeepGlobe), Table 5 evaluates visual grounding on RSVG, and Fig. 7 evaluates zero-shot classification on SATIN, all of which are outside RemoteSAM-270K and provide independent support for the model's generalization. The mask-to-bbox conversion in Eq. 2 and the counting/classification post-processing are algorithmic transformations, not circular derivations. The self-citations (e.g., RemoteCLIP for M2B, EPOC from a co-author, PoCa captioning) are implementation details and are not load-bearing for the central claim. There is no self-citation chain that forces the architecture; the main problem is benchmark overlap and hyperparameter selection on evaluation data. Hence the score is 6: several headline predictions reduce to training or tuning inputs, but meaningful independent evaluation exists elsewhere, so the paper is not wholly circular.

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

The central trainable claim rests on the quality of VLM-generated expressions and expert pseudo-labels, and the task unification rests on the sufficiency of masks as an output representation. The numeric results also depend on several thresholds and balance factors, most of which are either unreported or tuned on the evaluation benchmarks.

free parameters (4)
  • lambda (classification balance factor) = 0.5 (multi-label), 1.0 (image classification)
    Tuned on DIOR and UCM in Fig. 22 to maximize reported accuracy.
  • tau_cls (classification threshold) = 0.5 (default)
    Eq. 4; threshold for declaring positive labels, default 0.5, no sensitivity analysis.
  • tau_seg (semantic segmentation threshold) = not reported
    Eq. 1; threshold on category confidence P(c|I) for mask aggregation, value not specified.
  • SigLIP2 similarity cutoff for data filtering = not reported
    Section 3.1.2; threshold used to remove low-similarity pseudo-labels, not quantified.
assumptions (4)
  • domain assumption Pseudo-labels from expert models (GroundedSAM2, RMSIN, etc.) are accurate enough to serve as ground truth for training.
    Section 3.1.2 Masks Production; validated only on a 100-image sample (mIoU 74.14%, Appendix C Table 15).
  • domain assumption Qwen2-VL-generated referring expressions are semantically correct and aligned with the masks.
    Section 3.1.2 Expressions Creation; relies on the prior VLM's comprehension, no human verification of expression-mask alignment beyond similarity filtering.
  • domain assumption A pixel-level mask is a sufficient output representation from which region and image level task labels can be faithfully derived.
    Section 4; the entire task unification scheme assumes bounding boxes, counts, and labels can be recovered from masks without loss.
  • ad hoc to paper RemoteSAM-270K category coverage transfers to unseen remote sensing domains.
    The zero-shot results on Vaihingen/SATIN are presented as evidence, but the transfer itself is an assumption for other downstream uses.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RemoteSAM: Towards Segment Anything for Earth Observation." pith.science (2026). https://pith.science/paper/6JCUHFYM

@misc{pith2026250518022,
  author       = {Pith},
  title        = {Pith review of: RemoteSAM: Towards Segment Anything for Earth Observation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6JCUHFYM}},
  note         = {Machine review of arXiv:2505.18022}
}
read the original abstract

We aim to develop a robust yet flexible visual foundation model for Earth observation. It should possess strong capabilities in recognizing and localizing diverse visual targets while providing compatibility with various input-output interfaces required across different task scenarios. Current systems cannot meet these requirements, as they typically utilize task-specific architecture trained on narrow data domains with limited semantic coverage. Our study addresses these limitations from two aspects: data and modeling. We first introduce an automatic data engine that enjoys significantly better scalability compared to previous human annotation or rule-based approaches. It has enabled us to create the largest dataset of its kind to date, comprising 270K image-text-mask triplets covering an unprecedented range of diverse semantic categories and attribute specifications. Based on this data foundation, we further propose a task unification paradigm that centers around referring expression segmentation. It effectively handles a wide range of vision-centric perception tasks, including classification, detection, segmentation, grounding, etc, using a single model without any task-specific heads. Combining these innovations on data and modeling, we present RemoteSAM, a foundation model that establishes new SoTA on several earth observation perception benchmarks, outperforming other foundation models such as Falcon, GeoChat, and LHRS-Bot with significantly higher efficiency. Models and data are publicly available at https://github.com/1e12Leon/RemoteSAM.

Figures

Figures reproduced from arXiv: 2505.18022 by the authors.

Figure 1
Figure 1. Comparison of various foundation models for Earth ob [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Comparison of different foundation models of remote sensing. (a) Task-specific head methods [ [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. The composition of our proposed RemoteSAM-270K dataset. It is a generalized referring segmentation dataset, which is con [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (21 more)
Figure 5
Figure 5. Figure 5: Comparison between VLMs (ours) and rule-based [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 4
Figure 4. Figure 4: Comparison of semantic coverage on ours and other [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 6
Figure 6. Figure 6: Overview of our proposed RemoteSAM. It is a foundational model centered around a referring expression segmentation frame [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Comparison of zero-shot classification on SATIN with [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Inference examples of RemoteSAM on 8 visual-centric tasks. [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Detailed comparison with GroundedSAM2 on UCM [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Detailed comparison with GroundedSAM2 on AID [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]
Figure 11
Figure 11. Figure 11: Detailed comparison with GroundedSAM2 on MultiScene in SATIN Complex Scenes Task [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]
Figure 12
Figure 12. Figure 12: Detailed comparison with GroundedSAM2 on MLRSNet in SATIN Complex Scenes Task (Part I) [PITH_FULL_IMAGE:figures/full_fig_p018_12.png]
Figure 13
Figure 13. Figure 13: Detailed comparison with GroundedSAM2 on MLRSNet in SATIN Complex Scenes Task (Part II) [PITH_FULL_IMAGE:figures/full_fig_p019_13.png]
Figure 14
Figure 14. Figure 14: Task1: Referring Segmentation [PITH_FULL_IMAGE:figures/full_fig_p020_14.png]
Figure 15
Figure 15. Figure 15: Task2: Semantic Segmentation [PITH_FULL_IMAGE:figures/full_fig_p021_15.png]
Figure 16
Figure 16. Figure 16: Task3: Object Detection [PITH_FULL_IMAGE:figures/full_fig_p022_16.png]
Figure 17
Figure 17. Figure 17: Task4: Object Counting [PITH_FULL_IMAGE:figures/full_fig_p023_17.png]
Figure 18
Figure 18. Figure 18: Task5: Visual Grounding [PITH_FULL_IMAGE:figures/full_fig_p024_18.png]
Figure 19
Figure 19. Figure 19: Task6: Multi-Label Classification [PITH_FULL_IMAGE:figures/full_fig_p025_19.png]
Figure 20
Figure 20. Figure 20: Task7: Image Classification [PITH_FULL_IMAGE:figures/full_fig_p026_20.png]
Figure 21
Figure 21. Figure 21: Task8: Image Caption [PITH_FULL_IMAGE:figures/full_fig_p027_21.png]
Figure 22
Figure 22. Figure 22: Balance factor analysis of classification [PITH_FULL_IMAGE:figures/full_fig_p029_22.png]
Figure 23
Figure 23. Figure 23: Qualitative examples of EPOC refinement in Object Detection [PITH_FULL_IMAGE:figures/full_fig_p029_23.png]
Figure 24
Figure 24. Figure 24: Examples and categories of RemoteSAM-270k [PITH_FULL_IMAGE:figures/full_fig_p030_24.png]

Discussion (0). Sign in 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. PixDLM: A Dual-Path Multimodal Language Model for UAV Reasoning Segmentation

    cs.CV 2026-04 unverdicted novelty 7.0 of 10

    The work introduces the UAV Reasoning Segmentation task, the DRSeg benchmark dataset, and PixDLM as a baseline dual-path multimodal language model for reasoning-based segmentation in aerial imagery.

Reference graph

Works this paper leans on

88 extracted references · 67 canonical work pages · cited by 1 Pith paper

  1. [1]

    Fmars: Annotating remote sensing images for disas- ter management using foundation models

    Edoardo Arnaudo, Jacopo Lungo Vaschetti, Lorenzo Inno- centi, Luca Barco, Davide Lisi, Vanina Fissore, and Claudio Rossi. Fmars: Annotating remote sensing images for disas- ter management using foundation models. InIEEE IGARSS, pages 3920–3924. IEEE, 2024. 1

    2024

  2. [2]

    Foundation models defining a new era in vision: a survey and outlook

    Muhammad Awais, Muzammal Naseer, Salman Khan, Rao Muhammad Anwer, Hisham Cholakkal, Mubarak Shah, Ming-Hsuan Yang, and Fahad Shahbaz Khan. Foundation models defining a new era in vision: a survey and outlook. IEEE TPAMI, 2025. 1, 3

    2025

  3. [3]

    Grounding everything: Emerging localiza- tion properties in vision-language transformers

    Walid Bousselham, Felix Petersen, Vittorio Ferrari, and Hilde Kuehne. Grounding everything: Emerging localiza- tion properties in vision-language transformers. InCVPR, page 3828–3837. IEEE, 2024. 7

    2024

  4. [4]

    Learned embed- ding fields for multi-source, multi-temporal earth observa- tion imagery

    Christopher Brown, Michal Kazmierski, William Rucklidge, Valerie Pasquarella, and Evan Shelhamer. Learned embed- ding fields for multi-source, multi-temporal earth observa- tion imagery. InICLR Workshop on Machine Learning for Remote Sensing (ML4RS), 2024. 1

    2024

  5. [5]

    A billion-scale foundation model for remote sensing images

    Keumgang Cha, Junghoon Seo, and Taekyung Lee. A billion-scale foundation model for remote sensing images. IEEE JSTARS, page 1–17, 2024. 3

    2024

  6. [6]

    What makes for good image captions?, 2024

    Delong Chen, Samuel Cahyawijaya, Etsuko Ishii, Ho Shu Chan, Yejin Bang, and Pascale Fung. What makes for good image captions?, 2024. 4

    2024

  7. [7]

    Subobject-level image tokeniza- tion.arXiv preprint arXiv:2402.14327, 2024

    Delong Chen, Samuel Cahyawijaya, Jianfeng Liu, Baoyuan Wang, and Pascale Fung. Subobject-level image tokeniza- tion.arXiv preprint arXiv:2402.14327, 2024. 6

    arXiv 2024

  8. [8]

    Remote sensing of diverse urban environments: From the single city to multiple cities.RSE, 305:114108,

    Gang Chen, Yuyu Zhou, James A V oogt, and Eleanor C Stokes. Remote sensing of diverse urban environments: From the single city to multiple cities.RSE, 305:114108,

Show all 88 references
  1. [9]

    Fleet, and Ge- offrey Hinton

    Ting Chen, Saurabh Saxena, Lala Li, David J. Fleet, and Ge- offrey Hinton. Pix2seq: A language modeling framework for object detection, 2022. 2

    2022

  2. [10]

    Fleet, and Geoffrey Hinton

    Ting Chen, Saurabh Saxena, Lala Li, Tsung-Yi Lin, David J. Fleet, and Geoffrey Hinton. A unified sequence interface for vision tasks, 2022. 2

    2022

  3. [11]

    Mask grounding for referring image seg- mentation

    Yong Xien Chng, Henry Zheng, Yizeng Han, Xuchong Qiu, and Gao Huang. Mask grounding for referring image seg- mentation. InCVPR, pages 26573–26583, 2024. 2, 3

    2024

  4. [12]

    Cross-aware early fusion with stage-divided vision and language trans- former encoders for referring image segmentation.IEEE TMM, 26:5823–5833, 2023

    Yubin Cho, Hyunwoo Yu, and Suk-Ju Kang. Cross-aware early fusion with stage-divided vision and language trans- former encoders for referring image segmentation.IEEE TMM, 26:5823–5833, 2023. 8

    2023

  5. [13]

    Functional map of the world

    Gordon Christie, Neil Fendley, James Wilson, and Ryan Mukherjee. Functional map of the world. InCVPR, pages 6172–6180, 2018. 8

    2018

  6. [14]

    Satmae: Pre-training transformers for tem- poral and multi-spectral satellite imagery.NeurIPS, 35:197– 211, 2022

    Yezhen Cong, Samar Khanna, Chenlin Meng, Patrick Liu, Erik Rozi, Yutong He, Marshall Burke, David Lobell, and Stefano Ermon. Satmae: Pre-training transformers for tem- poral and multi-spectral satellite imagery.NeurIPS, 35:197– 211, 2022. 2, 3, 8

    2022

  7. [15]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. InCVPR, pages 248–255. Ieee, 2009. 8

    2009

  8. [16]

    Bert: Pre-training of deep bidirectional trans- formers for language understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional trans- formers for language understanding. InNAACL-HLT, pages 4171–4186, 2019. 7

    2019

  9. [17]

    Cross-modal bidirectional interaction model for referring remote sensing image segmentation.arXiv preprint arXiv:2410.08613, 2024

    Zhe Dong, Yuzhe Sun, Yanfeng Gu, and Tianzhu Liu. Cross-modal bidirectional interaction model for referring remote sensing image segmentation.arXiv preprint arXiv:2410.08613, 2024. 1, 3, 4, 5, 7, 8

    2024 arXiv

  10. [18]

    Minicpm: Unveiling the potential of small language models with scalable training strategies.arXiv preprint arXiv:2404.06395, 2024

    Shengding Hu, Yuge Tu, Xu Han, Chaoqun He, Ganqu Cui, Xiang Long, Zhi Zheng, Yewei Fang, Yuxiang Huang, Weilin Zhao, et al. Minicpm: Unveiling the potential of small language models with scalable training strategies.arXiv preprint arXiv:2404.06395, 2024. 8, 9

    2024 arXiv

  11. [19]

    Beyond one-to-one: Re- thinking the referring image segmentation

    Yutao Hu, Qixiong Wang, Wenqi Shao, Enze Xie, Zhenguo Li, Jungong Han, and Ping Luo. Beyond one-to-one: Re- thinking the referring image segmentation. InICCV, pages 4067–4077, 2023. 2, 4, 5

    2023

  12. [20]

    Bi-directional relationship inferring network for referring image segmentation

    Zhiwei Hu, Guang Feng, Jiayu Sun, Lihe Zhang, and Huchuan Lu. Bi-directional relationship inferring network for referring image segmentation. InCVPR, pages 4424– 4433, 2020. 8

    2020

  13. [21]

    Look before you leap: Learning landmark features for one-stage visual grounding

    Binbin Huang, Dongze Lian, Weixin Luo, and Shenghua Gao. Look before you leap: Learning landmark features for one-stage visual grounding. InCVPR, pages 16888–16897,

  14. [22]

    Referring im- age segmentation via cross-modal progressive comprehen- sion

    Shaofei Huang, Tianrui Hui, Si Liu, Guanbin Li, Yunchao Wei, Jizhong Han, Luoqi Liu, and Bo Li. Referring im- age segmentation via cross-modal progressive comprehen- sion. InCVPR, pages 10488–10497, 2020. 8

    2020

  15. [23]

    Linguistic structure guided context modeling for referring image segmentation

    Tianrui Hui, Si Liu, Shaofei Huang, Guanbin Li, Sansi Yu, Faxi Zhang, and Jizhong Han. Linguistic structure guided context modeling for referring image segmentation. InECCV, pages 59–75. Springer, 2020. 8

    2020

  16. [24]

    A survey of methods for addressing the chal- lenges of referring image segmentation.Neurocomputing, 583:127599, 2024

    Lixia Ji, Yunlong Du, Yiping Dang, Wenzhao Gao, and Han Zhang. A survey of methods for addressing the chal- lenges of referring image segmentation.Neurocomputing, 583:127599, 2024. 2, 3

    2024

  17. [25]

    Geochat: Grounded large vision-language model for remote sensing

    Kartik Kuckreja, Muhammad Sohail Danish, Muzammal Naseer, Abhijit Das, Salman Khan, and Fahad Shahbaz Khan. Geochat: Grounded large vision-language model for remote sensing. InCVPR, pages 27831–27840, 2024. 1, 2, 3, 7, 8, 9

    2024

  18. [26]

    Clearclip: Decom- posing clip representations for dense vision-language infer- ence

    Mengcheng Lan, Chaofeng Chen, Yiping Ke, Xinjiang Wang, Litong Feng, and Wayne Zhang. Clearclip: Decom- posing clip representations for dense vision-language infer- ence. InECCV, pages 143–160. Springer, 2024. 7

    2024

  19. [27]

    Exploring fine-grained image-text alignment for referring remote sensing image segmentation

    Sen Lei, Xinyu Xiao, Tianlin Zhang, Heng-Chao Li, Zhen- wei Shi, and Qing Zhu. Exploring fine-grained image-text alignment for referring remote sensing image segmentation. IEEE TGRS, 2024. 3, 7

    2024

  20. [28]

    Object detection in optical remote sensing images: A survey and a new benchmark.ISPRS PRS, 159:296–307,

    Ke Li, Gang Wan, Gong Cheng, Liqiu Meng, and Junwei Han. Object detection in optical remote sensing images: A survey and a new benchmark.ISPRS PRS, 159:296–307,

  21. [29]

    Segearth-ov: Towards training-free open-vocabulary segmentation for remote sens- ing images.arXiv preprint arXiv:2410.01768, 2024

    Kaiyu Li, ruixun Liu, Xiangyong Cao, Xueru Bai, Feng Zhou, Deyu Meng, and Zhi Wang. Segearth-ov: Towards training-free open-vocabulary segmentation for remote sens- ing images.arXiv preprint arXiv:2410.01768, 2024. 7

    2024 arXiv

  22. [30]

    Referring transformer: A one- step approach to multi-task visual grounding.NeurIPS, 34: 19652–19664, 2021

    Muchen Li and Leonid Sigal. Referring transformer: A one- step approach to multi-task visual grounding.NeurIPS, 34: 19652–19664, 2021. 8

    2021

  23. [31]

    Co- training transformer for remote sensing image classification, segmentation and detection.IEEE TGRS, 62:1–18, 2024

    Qingyun Li, Yushi Chen, Xin He, and Lingbo Huang. Co- training transformer for remote sensing image classification, segmentation and detection.IEEE TGRS, 62:1–18, 2024. 8

    2024

  24. [32]

    Toward open vocabulary aerial object detection with clip-activated student-teacher learning

    Yan Li, Weiwei Guo, Xue Yang, Ning Liao, Dunyun He, Jiaqi Zhou, and Wenxian Yu. Toward open vocabulary aerial object detection with clip-activated student-teacher learning. InECCV, pages 431–448. Springer, 2024. 1

    2024

  25. [33]

    Masked angle-aware autoen- coder for remote sensing images

    Zhihao Li, Biao Hou, Siteng Ma, Zitong Wu, Xianpeng Guo, Bo Ren, and Licheng Jiao. Masked angle-aware autoen- coder for remote sensing images. InECCV, pages 260–278. Springer, 2024. 2, 3, 7, 8

    2024

  26. [34]

    A real-time cross-modality correlation fil- tering method for referring expression comprehension

    Yue Liao, Si Liu, Guanbin Li, Fei Wang, Yanjie Chen, Chen Qian, and Bo Li. A real-time cross-modality correlation fil- tering method for referring expression comprehension. In CVPR, pages 10880–10889, 2020. 8

    2020

  27. [35]

    Progressive language-customized visual feature learn- ing for one-stage visual grounding.IEEE TIP, 31:4266– 4277, 2022

    Yue Liao, Aixi Zhang, Zhiyuan Chen, Tianrui Hui, and Si Liu. Progressive language-customized visual feature learn- ing for one-stage visual grounding.IEEE TIP, 31:4266– 4277, 2022. 8

    2022

  28. [36]

    Sphinx: The joint mixing of weights, tasks, and visual embeddings for multi-modal large language models.arXiv preprint arXiv:2311.07575, 2023

    Ziyi Lin, Chris Liu, Renrui Zhang, Peng Gao, Longtian Qiu, Han Xiao, Han Qiu, Chen Lin, Wenqi Shao, Keqin Chen, et al. Sphinx: The joint mixing of weights, tasks, and visual embeddings for multi-modal large language models.arXiv preprint arXiv:2311.07575, 2023. 8, 9

    2023 arXiv

  29. [37]

    Gres: Gen- eralized referring expression segmentation

    Chang Liu, Henghui Ding, and Xudong Jiang. Gres: Gen- eralized referring expression segmentation. InCVPR, pages 23592–23601, 2023. 4

    2023

  30. [38]

    Re- moteclip: A vision language foundation model for remote sensing.IEEE TGRS, 2024

    Fan Liu, Delong Chen, Zhangqingyun Guan, Xiaocong Zhou, Jiale Zhu, Qiaolin Ye, Liyong Fu, and Jun Zhou. Re- moteclip: A vision language foundation model for remote sensing.IEEE TGRS, 2024. 1, 6, 7, 8

    2024

  31. [39]

    Scale-invariant feature disentan- glement via adversarial learning for uav-based object detec- tion.arXiv preprint arXiv:2405.15465, 2024

    Fan Liu, Liang Yao, Chuanyi Zhang, Ting Wu, Xinlei Zhang, Xiruo Jiang, and Jun Zhou. Scale-invariant feature disentan- glement via adversarial learning for uav-based object detec- tion.arXiv preprint arXiv:2405.15465, 2024. 3

    2024 arXiv

  32. [40]

    Few-shot adaptation of multi-modal foundation models: A survey.Ar- tificial Intelligence Review, 57(10):268, 2024

    Fan Liu, Tianshu Zhang, Wenwen Dai, Chuanyi Zhang, Wenwen Cai, Xiaocong Zhou, and Delong Chen. Few-shot adaptation of multi-modal foundation models: A survey.Ar- tificial Intelligence Review, 57(10):268, 2024. 3

    2024

  33. [41]

    Visual instruction tuning, 2023

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning, 2023. 8, 9

    2023

  34. [42]

    Cross-modal progressive comprehension for referring segmentation.IEEE TPAMI, 44(9):4761–4775,

    Si Liu, Tianrui Hui, Shaofei Huang, Yunchao Wei, Bo Li, and Guanbin Li. Cross-modal progressive comprehension for referring segmentation.IEEE TPAMI, 44(9):4761–4775,

  35. [43]

    Rotated multi-scale interaction network for referring remote sensing image seg- mentation

    Sihan Liu, Yiwei Ma, Xiaoqing Zhang, Haowei Wang, Ji- ayi Ji, Xiaoshuai Sun, and Rongrong Ji. Rotated multi-scale interaction network for referring remote sensing image seg- mentation. InCVPR, pages 26658–26668, 2024. 2, 3, 4, 5, 7, 8

    2024

  36. [44]

    Caris: Context-aware re- ferring image segmentation

    Sun-Ao Liu, Yiheng Zhang, Zhaofan Qiu, Hongtao Xie, Yongdong Zhang, and Ting Yao. Caris: Context-aware re- ferring image segmentation. InACM MM, pages 779–788,

  37. [45]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In ICCV, pages 10012–10022, 2021. 7

    2021

  38. [46]

    On creating benchmark dataset for aerial image interpretation: Reviews, guidances, and million-aid.IEEE JSTARS, 14:4205–4230, 2021

    Yang Long, Gui-Song Xia, Shengyang Li, Wen Yang, Michael Ying Yang, Xiao Xiang Zhu, Liangpei Zhang, and Deren Li. On creating benchmark dataset for aerial image interpretation: Reviews, guidances, and million-aid.IEEE JSTARS, 14:4205–4230, 2021. 8

    2021

  39. [47]

    Change- aware sampling and contrastive learning for satellite images

    Utkarsh Mall, Bharath Hariharan, and Kavita Bala. Change- aware sampling and contrastive learning for satellite images. InCVPR, pages 5261–5270, 2023. 8

    2023

  40. [48]

    Seasonal contrast: Un- supervised pre-training from uncurated remote sensing data

    Oscar Manas, Alexandre Lacoste, Xavier Gir ´o-i Nieto, David Vazquez, and Pau Rodriguez. Seasonal contrast: Un- supervised pre-training from uncurated remote sensing data. InICCV, pages 9414–9423, 2021. 8

    2021

  41. [49]

    Prompting directsam for semantic contour extraction in remote sensing images

    Shiyu Miao, Delong Chen, Fan Liu, Chuanyi Zhang, Yanhui Gu, Shengjie Guo, and Jun Zhou. Prompting directsam for semantic contour extraction in remote sensing images. In ICASSP, pages 1–5. IEEE, 2025. 3

    2025

  42. [50]

    Lhrs-bot: Empowering remote sensing with vgi-enhanced large multimodal language model

    Dilxat Muhtar, Zhenshi Li, Feng Gu, Xueliang Zhang, and Pengfeng Xiao. Lhrs-bot: Empowering remote sensing with vgi-enhanced large multimodal language model. InECCV, pages 440–457. Springer, 2024. 1, 2, 3, 7, 8, 9

    2024

  43. [51]

    Rethinking transformers pre-training for multi- spectral satellite imagery

    Mubashir Noman, Muzammal Naseer, Hisham Cholakkal, Rao Muhammad Anwar, Salman Khan, and Fahad Shah- baz Khan. Rethinking transformers pre-training for multi- spectral satellite imagery. InCVPR, pages 27811–27819,

  44. [52]

    Sentinel-2 data for land cover/use mapping: A review.Re- mote sensing, 12(14):2291, 2020

    Darius Phiri, Matamyo Simwanda, Serajis Salekin, Vin- cent R Nyirenda, Yuji Murayama, and Manjula Ranagalage. Sentinel-2 data for land cover/use mapping: A review.Re- mote sensing, 12(14):2291, 2020. 8

    2020

  45. [53]

    Scale-mae: A scale-aware masked autoencoder for multiscale geospatial representation learning

    Colorado J Reed, Ritwik Gupta, Shufan Li, Sarah Brock- man, Christopher Funk, Brian Clipp, Kurt Keutzer, Salvatore Candido, Matt Uyttendaele, and Trevor Darrell. Scale-mae: A scale-aware masked autoencoder for multiscale geospatial representation learning. InICCV, pages 4065–4...

    2023

  46. [54]

    Scale-mae: A scale-aware masked autoencoder for multiscale geospatial representation learning

    Colorado J Reed, Ritwik Gupta, Shufan Li, Sarah Brock- man, Christopher Funk, Brian Clipp, Kurt Keutzer, Salvatore Candido, Matt Uyttendaele, and Trevor Darrell. Scale-mae: A scale-aware masked autoencoder for multiscale geospatial representation learning. InICCV, pages 4088–4...

    2023

  47. [55]

    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,

  48. [56]

    Satin: A multi-task metadataset for classifying satellite imagery using vision-language models.arXiv preprint arXiv:2304.11619,

    Jonathan Roberts, Kai Han, and Samuel Albanie. Satin: A multi-task metadataset for classifying satellite imagery using vision-language models.arXiv preprint arXiv:2304.11619,

    arXiv

  49. [57]

    Cus- tomized sam 2 for referring remote sensing image segmenta- tion, 2025

    Fu Rong, Meng Lan, Qian Zhang, and Lefei Zhang. Cus- tomized sam 2 for referring remote sensing image segmenta- tion, 2025. 8

    2025

  50. [58]

    Zero-shot ground- ing of objects from natural language queries

    Arka Sadhu, Kan Chen, and Ram Nevatia. Zero-shot ground- ing of objects from natural language queries. InICCV, pages 4694–4703, 2019. 8

    2019

  51. [59]

    Ringmo: A remote sensing foundation model with masked image modeling.IEEE TGRS, 61:1–22, 2022

    Xian Sun, Peijin Wang, Wanxuan Lu, Zicong Zhu, Xiao- nan Lu, Qibin He, Junxi Li, Xuee Rong, Zhujun Yang, Hao Chang, et al. Ringmo: A remote sensing foundation model with masked image modeling.IEEE TGRS, 61:1–22, 2022. 1, 7, 8

    2022

  52. [60]

    Ringmo: A remote sensing foundation model with masked image modeling.IEEE TGRS, 61:1–22, 2023

    Xian Sun, Peijin Wang, Wanxuan Lu, Zicong Zhu, Xiao- nan Lu, Qibin He, Junxi Li, Xuee Rong, Zhujun Yang, Hao Chang, Qinglin He, Guang Yang, Ruiping Wang, Jiwen Lu, and Kun Fu. Ringmo: A remote sensing foundation model with masked image modeling.IEEE TGRS, 61:1–22, 2023. 3

    2023

  53. [61]

    Siglip 2: Multilingual vision- language encoders with improved semantic understand- ing, localization, and dense features.arXiv preprint arXiv:2502.14786, 2025

    Michael Tschannen, Alexey Gritsenko, Xiao Wang, Muham- mad Ferjad Naeem, Ibrahim Alabdulmohsin, Nikhil Parthasarathy, Talfan Evans, Lucas Beyer, Ye Xia, Basil Mustafa, Olivier H ´enaff, Jeremiah Harmsen, Andreas Steiner, and Xiaohua Zhai. Siglip 2: Multilingual vision- languag...

    2025 arXiv

  54. [62]

    Arti- ficial intelligence to advance earth observation: A review of models, recent trends, and pathways forward.IEEE GRSM,

    Devis Tuia, Konrad Schindler, Beg ¨um Demir, Xiao Xiang Zhu, Mrinalini Kochupillai, Sa ˇso Dˇzeroski, Jan N van Rijn, Holger H Hoos, Fabio Del Frate, Mihai Datcu, et al. Arti- ficial intelligence to advance earth observation: A review of models, recent trends, and pathways for...

  55. [63]

    Remote sensing for agriculture in the era of industry 5.0–a survey.IEEE JSTARS, 2024

    Nancy Victor, Praveen Kumar Reddy Maddikunta, Delphin Raj Kesari Mary, Ramalingam Murugan, Rajeswari Chengo- den, Thippa Reddy Gadekallu, Nitin Rakesh, Yaodong Zhu, and Jeongyeup Paek. Remote sensing for agriculture in the era of industry 5.0–a survey.IEEE JSTARS, 2024. 1

    2024

  56. [64]

    Advancing plain vision transformer toward remote sensing foundation model.IEEE TGRS, 61:1–15, 2022

    Di Wang, Qiming Zhang, Yufei Xu, Jing Zhang, Bo Du, Dacheng Tao, and Liangpei Zhang. Advancing plain vision transformer toward remote sensing foundation model.IEEE TGRS, 61:1–15, 2022. 8

    2022

  57. [65]

    Samrs: Scaling-up re- mote sensing segmentation dataset with segment anything model.NeurIPS, 36:8815–8827, 2023

    Di Wang, Jing Zhang, Bo Du, Minqiang Xu, Lin Liu, Dacheng Tao, and Liangpei Zhang. Samrs: Scaling-up re- mote sensing segmentation dataset with segment anything model.NeurIPS, 36:8815–8827, 2023. 4, 8

    2023

  58. [66]

    Sclip: Rethink- ing self-attention for dense vision-language inference.arXiv preprint arXiv:2312.01597, 2023

    Feng Wang, Jieru Mei, and Alan Yuille. Sclip: Rethink- ing self-attention for dense vision-language inference.arXiv preprint arXiv:2312.01597, 2023. 7

    2023 arXiv

  59. [67]

    Loveda: A remote sensing land-cover dataset for domain adaptive semantic segmentation.arXiv preprint arXiv:2110.08733, 2021

    Junjue Wang, Zhuo Zheng, Ailong Ma, Xiaoyan Lu, and Yanfei Zhong. Loveda: A remote sensing land-cover dataset for domain adaptive semantic segmentation.arXiv preprint arXiv:2110.08733, 2021. 4

    2021 arXiv

  60. [68]

    Ofa: Unifying architectures, tasks, and modalities through a simple sequence-to-sequence learning framework

    Peng Wang, An Yang, Rui Men, Junyang Lin, Shuai Bai, Zhikang Li, Jianxin Ma, Chang Zhou, Jingren Zhou, and Hongxia Yang. Ofa: Unifying architectures, tasks, and modalities through a simple sequence-to-sequence learning framework. InICML, pages 23318–23340. PMLR, 2022. 2

    2022

  61. [69]

    Ssl4eo- s12: A large-scale multimodal, multitemporal dataset for self-supervised learning in earth observation [software and data sets].IEEE GRSM, 11(3):98–106, 2023

    Yi Wang, Nassim Ait Ali Braham, Zhitong Xiong, Cheny- ing Liu, Conrad M Albrecht, and Xiao Xiang Zhu. Ssl4eo- s12: A large-scale multimodal, multitemporal dataset for self-supervised learning in earth observation [software and data sets].IEEE GRSM, 11(3):98–106, 2023. 8

    2023

  62. [70]

    isaid: A large- scale dataset for instance segmentation in aerial images

    Syed Waqas Zamir, Aditya Arora, Akshita Gupta, Salman Khan, Guolei Sun, Fahad Shahbaz Khan, Fan Zhu, Ling Shao, Gui-Song Xia, and Xiang Bai. isaid: A large- scale dataset for instance segmentation in aerial images. In CVPRW, pages 28–37, 2019. 4, 7

    2019

  63. [71]

    Towards robust referring image seg- mentation.IEEE TIP, 2024

    Jianzong Wu, Xiangtai Li, Xia Li, Henghui Ding, Yunhai Tong, and Dacheng Tao. Towards robust referring image seg- mentation.IEEE TIP, 2024. 2, 8

    2024

  64. [72]

    Dota: A large-scale dataset for object detection in aerial images

    Gui-Song Xia, Xiang Bai, Jian Ding, Zhen Zhu, Serge Be- longie, Jiebo Luo, Mihai Datcu, Marcello Pelillo, and Liang- pei Zhang. Dota: A large-scale dataset for object detection in aerial images. InCVPR, pages 3974–3983, 2018. 4, 7

    2018

  65. [73]

    Qwen2.5 tech- nical report, 2025

    An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, and Chengyuan Li et al. Qwen2.5 tech- nical report, 2025. 2, 3, 4, 8

    2025

  66. [74]

    Graph adversarial self-supervised learning.NeurIPS, 34:14887– 14899, 2021

    Longqi Yang, Liangliang Zhang, and Wenjing Yang. Graph adversarial self-supervised learning.NeurIPS, 34:14887– 14899, 2021. 8

    2021

  67. [75]

    Lavt: Language-aware vi- sion transformer for referring image segmentation

    Zhao Yang, Jiaqi Wang, Yansong Tang, Kai Chen, Heng- shuang Zhao, and Philip HS Torr. Lavt: Language-aware vi- sion transformer for referring image segmentation. InCVPR, pages 18155–18165, 2022. 8

    2022

  68. [76]

    Falcon: A remote sensing vision-language foun- dation model.arXiv preprint arXiv:2503.11070, 2025

    Kelu Yao, Nuo Xu, Rong Yang, Yingying Xu, Zhuoyan Gao, Titinunt Kitrungrotsakul, Yi Ren, Pu Zhang, Jin Wang, Ning Wei, et al. Falcon: A remote sensing vision-language foun- dation model.arXiv preprint arXiv:2503.11070, 2025. 1, 2, 7, 8

    2025

  69. [77]

    Domain-invariant progressive knowledge distillation for uav-based object detection.IEEE GRSL, 2024

    Liang Yao, Fan Liu, Chuanyi Zhang, Zhiquan Ou, and Ting Wu. Domain-invariant progressive knowledge distillation for uav-based object detection.IEEE GRSL, 2024. 3

    2024

  70. [78]

    Shifting more attention to visual backbone: Query-modulated refinement networks for end-to-end visual grounding

    Jiabo Ye, Junfeng Tian, Ming Yan, Xiaoshan Yang, Xuwu Wang, Ji Zhang, Liang He, and Xin Lin. Shifting more attention to visual backbone: Query-modulated refinement networks for end-to-end visual grounding. InCVPR, pages 15502–15512, 2022. 8

    2022

  71. [79]

    Rrsis: Referring remote sensing image seg- mentation.IEEE TGRS, 2024

    Zhenghang Yuan, Lichao Mou, Yuansheng Hua, and Xiao Xiang Zhu. Rrsis: Referring remote sensing image seg- mentation.IEEE TGRS, 2024. 2, 3, 4, 5, 8

    2024

  72. [80]

    Rsvg: Exploring data and models for visual grounding on remote sensing data

    Yang Zhan, Zhitong Xiong, and Yuan Yuan. Rsvg: Exploring data and models for visual grounding on remote sensing data. IEEE TGRS, 61:1–13, 2023. 7, 8, 9

    2023

  73. [81]

    Skyeyegpt: Uni- fying remote sensing vision-language tasks via instruction tuning with large language model.ISPRS PRS, 221:64–77,

    Yang Zhan, Zhitong Xiong, and Yuan Yuan. Skyeyegpt: Uni- fying remote sensing vision-language tasks via instruction tuning with large language model.ISPRS PRS, 221:64–77,

  74. [82]

    Vision-language models for vision tasks: A survey.IEEE TPAMI, 2024

    Jingyi Zhang, Jiaxing Huang, Sheng Jin, and Shijian Lu. Vision-language models for vision tasks: A survey.IEEE TPAMI, 2024. 1

    2024

  75. [83]

    Earthgpt: A universal multi-modal large lan- guage model for multi-sensor image comprehension in re- mote sensing domain.IEEE TGRS, 2024

    Wei Zhang, Miaoxin Cai, Tong Zhang, Yin Zhuang, and Xuerui Mao. Earthgpt: A universal multi-modal large lan- guage model for multi-sensor image comprehension in re- mote sensing domain.IEEE TGRS, 2024. 2

    2024

  76. [84]

    Hierarchical and robust convolutional neural network for very high-resolution remote sensing object detection.IEEE TGRS, 57(8):5535–5548, 2019

    Yuanlin Zhang, Yuan Yuan, Yachuang Feng, and Xiaoqiang Lu. Hierarchical and robust convolutional neural network for very high-resolution remote sensing object detection.IEEE TGRS, 57(8):5535–5548, 2019. 4

    2019

  77. [85]

    Rs5m and georsclip: A large scale vision-language dataset and a large vision-language model for remote sensing.IEEE TGRS, 2024

    Zilun Zhang, Tiancheng Zhao, Yulong Guo, and Jianwei Yin. Rs5m and georsclip: A large scale vision-language dataset and a large vision-language model for remote sensing.IEEE TGRS, 2024. 1

    2024

  78. [86]

    Extract free dense labels from clip

    Chong Zhou, Chen Change Loy, and Bo Dai. Extract free dense labels from clip. InECCV, 2022. 7

    2022

  79. [87]

    Geoground: A unified large vision-language model

    Yue Zhou, Mengcheng Lan, Xiang Li, Yiping Ke, Xue Jiang, Litong Feng, and Wayne Zhang. Geoground: A unified large vision-language model. for remote sensing visual grounding. arXiv preprint arXiv:2411.11904, 2024. 7, 8

    2024 arXiv

  80. [88]

    aircraft

    Deyao Zhu, Jun Chen, Xiaoqian Shen, Xiang Li, and Mo- hamed Elhoseiny. Minigpt-4: Enhancing vision-language understanding with advanced large language models.arXiv preprint arXiv:2304.10592, 2023. 3, 8, 9 Appendix A. Quantitative comparison results for remaining tasks In this ...

    2023 arXiv

Pith tools

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