Pith. sign in

REVIEW 3 major objections 6 minor 71 references

This paper introduces spatio-temporal panoptic scene graph generation for satellite video, presents the large-scale T-STAR benchmark, and shows that its STCL framework predicts temporally grounded relationship triplets better than existing

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T-STAR provides 1.19M instance masks and 3.83M spatio-temporal relationship triplets from 150 Jilin-1 satellite videos, defining a new task of panoptic scene graph generation in satellite video.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection T-STAR is a genuinely new, large-scale benchmark for satellite-video scene graph prediction; the method is sound and the experiments are thorough, but the absence of annotation-quality metrics undercuts the benchmark's central claim, so accept conditional on that being addressed. the 3 major comments →

arxiv 2607.21228 v1 pith:S4BRJAMO submitted 2026-07-23 cs.CV

T-STAR: A Large-Scale Benchmark for Spatio-Temporal Panoptic Scene Graph Generation in Satellite Video

Linlin Wang , Xue Yang , Zhihuang Zhou , Zhenyu Zhong , Ruiyuan Zhang , Yansheng Li This is my paper
classification cs.CV
keywords satellite videopanoptic scene graph generationspatio-temporal relationship predictioninstance masksbenchmark datasetvideo understandingremote sensingtemporal reasoning
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 tries to establish that understanding satellite video can move beyond detecting, segmenting, and tracking objects to structured relational reasoning about dynamic scenes. It defines a task—spatio-temporal panoptic scene graph generation (TPSG)—where every scene element is given an identity-consistent mask and object pairs are annotated as triplets carrying explicit temporal spans, and it introduces T-STAR as the first large-scale benchmark for this task, with 1,188,193 instance masks and 3,832,449 triplets across 39 object and 70 relationship categories. It further claims that a dedicated model, STCL, which jointly handles cross-frame identity matching, spatial context among pairs, and multi-scale temporal evolution, clearly outperforms six existing relationship-prediction baselines under both evaluation settings. If the paper is right, the remote-sensing community gains a common evaluation ground for relational reasoning in satellite video, and a strong baseline to beat.

Core claim

On its own terms, the paper claims that satellite video is meaningfully described by identity-consistent instance masks plus spatio-temporally grounded relationship triplets <subject, relationship, object> with explicit temporal spans. To make that claim testable, it contributes T-STAR: 150 videos, 34,333 frames, 1,188,193 panoptic instance masks, and 3,832,449 triplets over 39 object and 70 relationship categories, annotated under a protocol requiring each triplet to be visually verifiable in its temporal interval. It then proposes the STCL framework—memory-guided matching for cross-frame identity, spatial context enhancement for pair-level context, multi-scale temporal learning for evoluti

What carries the argument

The central object is the temporally grounded triplet <M_S, r_Δt, M_O>: subject and object are identity-consistent mask trajectories, and r_Δt is a relationship category tagged with the temporal interval (possibly multiple disjoint intervals) over which it holds. The paper factorizes the task as P(G|V)=P(M|V)P(Y|M,V), separating video panoptic parsing from relationship prediction. The proposed STCL model carries the relationship stage: memory-guided matching (MGM) keeps a trajectory memory bank and greedily matches current masks to trajectories under category and recency constraints; spatial context enhancement (SCE) builds a graph whose nodes are instance pairs and propagates context by mes

Load-bearing premise

The load-bearing premise is that all 70 relationship categories, including functional and identity-discrimination relations such as 'collaborate with' and 'same group as', are visually verifiable from roughly one-meter satellite video within their annotated temporal spans, as asserted in Section III-C—yet the paper reports no inter-annotator agreement to check it.

What would settle it

Take a random sample of triplets spanning the seven relationship groups, have independent annotators re-label them from the same video intervals, and compute chance-corrected agreement; if agreement on functional and identity-discrimination categories is at chance, those ground-truth labels are not visually verifiable. A simpler version: show annotators the video interval without category names and see whether they can reliably sort 'collaborate with' versus 'not collaborate with'.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Satellite video understanding can be evaluated as a structured reasoning task, with a shared protocol (TR@K, mTR@K, FTR@K) and strict criteria requiring correct categories, temporal overlap, and mask overlap for a predicted triplet to count.
  • Trajectory quality gates relationship quality: the same STCL model roughly doubles its FTR when trajectories come from the stronger image-parsing-plus-tracking pipeline rather than video panoptic segmentation alone.
  • Relationship semantics in this domain are inherently temporal; categories such as approach, take off from, and docked at the same dock as cannot be judged from one frame, which is why local temporal aggregation baselines underperform.
  • Natural- and aerial-video scene graph methods do not transfer directly to meter-level satellite video, so T-STAR provides a needed testbed where they become baselines.
  • Long-tailed object and relationship distributions mean future work will need debiasing and rare-category handling, not just better backbones.

Where Pith is reading between the lines

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

  • Editorial inference: The same TPSG formulation could transfer to other dynamic overhead imagery, including drone and high-resolution aerial video, where objects are larger and appearance cues richer; T-STAR would then serve as a source-domain testbed for cross-domain relationship reasoning.
  • Editorial inference: Since the ablation singles out multi-scale temporal learning as the largest contributor, a concrete prediction is that future improvements on this benchmark will come from longer-range temporal architectures such as memory networks or transformers with explicit duration modeling, rather than stronger per-frame appearance features.
  • Editorial inference: If the visual-verifiability assumption holds, the benchmark opens a route toward open-world TPSG—recognizing relationship categories beyond the 70 defined ones—because the temporal-grounded triplet format is agnostic to the category list.
  • Editorial inference: A focused reliability study with independent re-annotation is the natural next step; without it, the 70-category taxonomy should be treated as agreed-by-protocol rather than measured-agreement ground truth.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper introduces spatio-temporal panoptic scene graph generation (TPSG) in satellite video as a new task, along with the T-STAR benchmark dataset: 150 Jilin-1 videos, roughly 34k frames, about 1.19M instance masks, and about 3.83M spatio-temporal triplets over 39 object and 70 relationship categories. The authors also propose STCL, a two-stage framework in which Stage I performs video panoptic parsing (IPS+T or VPS) and Stage II predicts triplets using memory-guided matching, spatial context enhancement, and multi-scale temporal learning. Experiments on T-STAR compare STCL against six baselines under PredCls and SGDet settings using TR/mTR/FTR metrics at K=200/500/1000, and report consistent gains for the proposed method, with ablations for the three components and temporal depth.

Significance. If the benchmark annotation is reliable, T-STAR would be a substantial new resource: it is, to the best of my knowledge, the first large-scale TPSG benchmark for satellite video, and the paper makes a clear case that existing image/video scene-graph datasets do not cover the satellite-video regime. The task formulation in Eq. (1) is reasonable, and the decision to use identity-consistent mask trajectories as graph nodes is well motivated for meter-level satellite imagery. The authors provide a clear comparison to prior satellite-video datasets and to natural-video SGG baselines, and they ship dataset and code URLs. The experimental framework is internally consistent, and the ablations are informative. The main weakness is that the central dataset claim depends on an annotation-reliability assertion for which no quantitative evidence is provided, and the experimental claims are made from single-seed runs without variance estimates.

major comments (3)
  1. [§III-B.2 and §III-C] The benchmark’s validity rests on the claim that every relationship is visually verifiable in meter-level satellite video, but the taxonomy includes identity-discrimination and functional predicates such as 'same group as', 'different group from', 'collaborate with', and 'co-storage with' that are not obviously inferable from ~1 m imagery. Section III-C states that each triplet must be visually verifiable but gives no operational definition of 'visually verifiable', no inter-annotator agreement statistics, and no per-category annotation-error analysis. The closed-loop workflow (four annotators, two experts) is described but not quantified. Without IAA or a validation sample, the 3.8M triplets may contain systematic label noise, which would directly affect all TR/mTR/FTR scores in Tables III–V. This is a load-bearing missing support rather than a stylistic issue; the authors should either
  2. [§V-A, Tables III–V] All experiments are single runs with a fixed random seed (42), and no standard deviations, confidence intervals, or multiple-seed results are reported. Some of the margins between STCL and the best baseline are not large in absolute terms—for example, Table IV IPS+T/SGDet/θt=0.5, FTR@1000 is 29.13 vs 23.49 for Transformer and 23.05 for IRG, but several TR@1000 values are within 2–4 points across methods. Without repeated runs, the claim that STCL 'substantially outperforms' all six baselines is not statistically grounded. I recommend reporting mean±std over at least three seeds for the main tables, and significance tests where relevant.
  3. [§V-A, §IV-C.1] The PredCls setting is under-specified with respect to the MGM module. Section V-A says PredCls provides ground-truth instance trajectories and categories, but MGM is designed to refine cross-frame instance association. If ground-truth trajectories are already identity-consistent, it is unclear what MGM operates on in PredCls, and whether it is active in the PredCls rows of Table III. The ablations in Table VI are only reported under SGDet, so the contribution of MGM in PredCls cannot be verified. Please clarify the exact inputs to MGM under each evaluation setting and, ideally, ablate MGM under PredCls as well.
minor comments (6)
  1. [Table III] Formatting issue: '16.0511.58' should read '16.05/11.58' (separator and spacing are missing).
  2. [§V-A] The temporal overlap criterion is described as a threshold but not defined precisely; please state whether it is temporal IoU, intersection-over-union of intervals, or another measure.
  3. [§III-B.1] The object taxonomy includes 'kelvin' and 'turbulent' as accompanying wake-like regions; these are treated as object categories in Fig. 4. A sentence clarifying that these are visual wake signatures, not discrete physical objects, would help readers unfamiliar with maritime imagery.
  4. [§III-D.1 and Fig. 3] The long-tail discussion is qualitative. Reporting the number of instances/triplets for the top and bottom categories, or an imbalance ratio, would strengthen the description and help future methods address the imbalance.
  5. [§III-A and Table II] The data source mentions both publicly accessible and commercially acquired Jilin-1 videos. Please clarify licensing and any restrictions on redistribution, since the dataset is promised for public release.
  6. [§II-B] The text refers to an 'AeroEye dataset' in aerial-video SGG, but the reference list only includes CYCLO [22]. Please add the AeroEye source or clarify that it is part of [22].

Circularity Check

0 steps flagged

No significant circularity: the benchmark and method claims are supported by dataset statistics and controlled experiments, not by self-citation or construction.

full rationale

T-STAR is a data-contribution and empirical-method paper; the central claims are that the dataset is the first large-scale TPSG benchmark in satellite video and that STCL outperforms six baselines. Neither claim is derived from a fitted constant or from a definition that contains the result: the dataset statistics (1,188,193 masks, 3,832,449 triplets, 39/70 categories) are reported from annotations, and the performance comparisons in Tables III–VII are standard controlled evaluations with shared Stage-I parsing and pair-selection, so STCL's gains do not reduce to its own training objective. The annotation protocol's statement that 'Each triplet is required to be visually verifiable within its corresponding temporal interval' and the lack of inter-annotator agreement metrics are data-quality/validity limitations, not circularity, because the benchmark's reliability is not used as the proof of a derived prediction. The authors' self-citations (e.g., STAR [60], prototype/bias-aware SGG works [63,64]) appear in Related Work as contextual prior art and are not invoked to justify the benchmark's validity, the taxonomy, or the method's effectiveness. Accordingly, there is no exhibited circular step.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 2 invented entities

The central claim rests on the dataset annotations being reliable and on the task abstraction being useful. The free parameters are hyperparameters of the proposed model; they do not undermine the dataset claim, but the annotation axioms — especially visual verifiability of all relationship categories — are load-bearing and unquantified.

free parameters (3)
  • loss weights λ1, λ2, λ3 = 1, 5, 1
    Hand-chosen weights for pair, predicate, and span losses (§IV-C4); no sensitivity analysis is provided.
  • TCN depth L and dilation schedule = L=4, dilation rates 1,2,4,8
    Selected via ablation on the dataset (§V-D); performance varies with L.
  • learning rate = 1e-4 (Adam)
    Standard optimization choice, not justified or swept.
axioms (4)
  • domain assumption The TPSG task formulation with spatio-temporal triplets is the right abstraction for structured satellite video understanding
    Motivated in §I but not empirically validated; no user study or downstream task demonstrates utility.
  • domain assumption All 70 relationship categories are visually verifiable from ~1m satellite video
    Stated in §III-C; no inter-annotator agreement evidence; several categories ('collaborate with', 'same group as') appear high-level.
  • domain assumption The evaluation protocol from PVSG [20] (temporal overlap + per-frame mask IoU) transfers to satellite video
    Adopted in §V-A without justification of thresholds.
  • standard math Factorization P(G|V)=P(M|V)P(Y|M,V)
    Valid by chain rule; used to justify the two-stage pipeline (§IV-A).
invented entities (2)
  • Identity-discrimination and functional relationship categories ('same group as', 'collaborate with', 'co-storage with') no independent evidence
    purpose: Extend spatio-temporal relationship semantics beyond geometric and motion descriptions to group-level and functional interactions.
    These labels are defined by the annotation taxonomy but lack an external falsifiable handle; they are not directly observable from pixel evidence at ~1m resolution.
  • Wake-like object categories 'kelvin' and 'turbulent' independent evidence
    purpose: Model ship wake regions as panoptic scene elements to capture maritime activity cues.
    Kelvin wakes are physically observable in satellite imagery, though detecting them as instances at 1m resolution is challenging.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of T-STAR: A Large-Scale Benchmark for Spatio-Temporal Panoptic Scene Graph Generation in Satellite Video." pith.science (2026). https://pith.science/paper/S4BRJAMO

@misc{pith2026260721228,
  author       = {Pith},
  title        = {Pith review of: T-STAR: A Large-Scale Benchmark for Spatio-Temporal Panoptic Scene Graph Generation in Satellite Video},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S4BRJAMO}},
  note         = {Machine review of arXiv:2607.21228}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Structured understanding of satellite video is essential for advancing dynamic geospatial scene analysis from low-level perception to high-level cognition. To move beyond object-centric perception, this paper introduces spatio-temporal panoptic scene graph generation (TPSG) in satellite video as a new benchmark task. TPSG aims to generate a structured graph composed of a set of triplets <subject, relationship, object> with explicit temporal spans, thereby describing dynamic geospatial scenes by jointly modeling identity-consistent instance masks and spatio-temporal relationships among panoptic scene elements. However, there is still no dedicated dataset for TPSG in satellite video. Moreover, TPSG in satellite video is intrinsically challenging, as objects are often small and weakly textured, cross-frame association is easily disrupted by occlusion and background clutter, and relationship semantics are highly coupled with spatial structure and temporal evolution. Consequently, TPSG models developed for natural videos are not directly applicable to satellite video. This paper presents T-STAR, a large-scale benchmark dataset for TPSG in satellite video, comprising over 1.1 million instance masks and over 3.8 million spatio-temporal triplets across 39 fine-grained object categories and 70 fine-grained relationship categories. To enable TPSG in satellite video, we propose a unified framework to enhance cross-frame instance consistency and spatio-temporal relationship prediction. Extensive experiments demonstrate the significance of T-STAR and the effectiveness of the proposed framework, establishing a strong benchmark for future research on structured satellite video understanding. The dataset and code are available at https://github.com/linlin-dev/T-STAR.

Figures

Figures reproduced from arXiv: 2607.21228 by Linlin Wang, Ruiyuan Zhang, Xue Yang, Yansheng Li, Zhenyu Zhong, Zhihuang Zhou.

Figure 1
Figure 1. Figure 1: Illustration of spatio-temporal panoptic scene graph generation (TPSG) in satellite video. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Representative annotated scenes from the T-STAR dataset. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Statistics and visualization of objects (a) and relationships (b) from the T-STAR dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Interaction mapping among subject categories, relationship categories, and object categories. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Overall framework for the TPSG task. Given a satellite video sequence, instance features are obtained via video [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Illustration of the spatial graph construction. [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Qualitative visualization of STCL in a port scene. [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative visualization of STCL in an airport scene. [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

71 extracted references · 1 linked inside Pith

  1. [1]

    Using satellite imagery to understand and promote sustainable development,

    M. Burke, A. Driscoll, D. B. Lobell, and S. Ermon, “Using satellite imagery to understand and promote sustainable development,”Science, vol. 371, no. 6535, p. eabe8628, 2021

    2021

  2. [2]

    A semantic-enhanced multi-modal remote sensing foundation model for earth observation,

    K. Wu, Y . Zhang, L. Ru, B. Dang, J. Lao, L. Yu, J. Luo, Z. Zhu, Y . Sun, J. Zhanget al., “A semantic-enhanced multi-modal remote sensing foundation model for earth observation,”Nature Machine Intelligence, vol. 7, no. 8, pp. 1235–1249, 2025

    2025

  3. [3]

    Vehicle perception from satellite,

    B. Zhao, P. Han, and X. Li, “Vehicle perception from satellite,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 4, pp. 2545–2554, 2024

    2024

  4. [4]

    Hazy low-quality satellite video restoration via learning optimal joint degradation patterns and continuous-scale super- resolution reconstruction,

    N. Ni and L. Zhang, “Hazy low-quality satellite video restoration via learning optimal joint degradation patterns and continuous-scale super- resolution reconstruction,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025, pp. 12 690–12 699

    2025

  5. [5]

    Highly efficient and unsupervised framework for moving object detection in satellite videos,

    C. Xiao, W. An, Y . Zhang, Z. Su, M. Li, W. Sheng, M. Pietik ¨ainen, and L. Liu, “Highly efficient and unsupervised framework for moving object detection in satellite videos,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 12, pp. 11 532–11 539, 2024

    2024

  6. [6]

    Moving vehicle detection for remote sensing video surveillance with nonstationary satellite platform,

    J. Zhang, X. Jia, J. Hu, and K. Tan, “Moving vehicle detection for remote sensing video surveillance with nonstationary satellite platform,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 9, pp. 5185–5198, 2022

    2022

  7. [7]

    Spatio-temporal dual-branch network with predictive feature learning for satellite video object seg- mentation,

    Y . Zhong, M. Shu, Z. Liu, and X. Lu, “Spatio-temporal dual-branch network with predictive feature learning for satellite video object seg- mentation,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–15, 2022

    2022

  8. [8]

    A multitask benchmark dataset for satellite video: Object de- tection, tracking, and segmentation,

    S. Li, Z. Zhou, M. Zhao, J. Yang, W. Guo, Y . Lv, L. Kou, H. Wang, and Y . Gu, “A multitask benchmark dataset for satellite video: Object de- tection, tracking, and segmentation,”IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–21, 2023

    2023

  9. [9]

    Mctracker: Satellite video multi- object tracking considering inter-frame motion correlation and multi- scale cascaded feature enhancement,

    B. Wang, H. Sui, G. Ma, and Y . Zhou, “Mctracker: Satellite video multi- object tracking considering inter-frame motion correlation and multi- scale cascaded feature enhancement,”ISPRS Journal of Photogrammetry and Remote Sensing, vol. 214, pp. 82–103, 2024

    2024

  10. [10]

    Satellite video single object tracking: A systematic review and an oriented object tracking benchmark,

    Y . Chen, Y . Tang, Y . Xiao, Q. Yuan, Y . Zhang, F. Liu, J. He, and L. Zhang, “Satellite video single object tracking: A systematic review and an oriented object tracking benchmark,”ISPRS Journal of Pho- togrammetry and Remote Sensing, vol. 210, pp. 212–240, 2024

    2024

  11. [11]

    Deep feature extraction and motion representation for satellite video scene classification,

    Y . Gu, H. Liu, T. Wang, S. Li, and G. Gao, “Deep feature extraction and motion representation for satellite video scene classification,”Science China Information Sciences, vol. 63, no. 4, p. 140307, 2020

    2020

  12. [12]

    Satellite video multi- label scene classification with spatial and temporal feature cooperative encoding: A benchmark dataset and method,

    W. Guo, S. Li, F. Chen, Y . Sun, and Y . Gu, “Satellite video multi- label scene classification with spatial and temporal feature cooperative encoding: A benchmark dataset and method,”IEEE Transactions on Image Processing, vol. 33, pp. 2238–2251, 2024

    2024

  13. [13]

    Spatial-temporal transformer for dynamic scene graph generation,

    Y . Cong, W. Liao, H. Ackermann, B. Rosenhahn, and M. Y . Yang, “Spatial-temporal transformer for dynamic scene graph generation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 16 372–16 382

    2021

  14. [14]

    Action genome: Actions as compositions of spatio-temporal scene graphs,

    J. Ji, R. Krishna, L. Fei-Fei, and J. C. Niebles, “Action genome: Actions as compositions of spatio-temporal scene graphs,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 10 236–10 247

    2020

  15. [15]

    A com- prehensive survey of scene graphs: Generation and application,

    X. Chang, P. Ren, P. Xu, Z. Li, X. Chen, and A. Hauptmann, “A com- prehensive survey of scene graphs: Generation and application,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 1, pp. 1–26, 2023

    2023

  16. [16]

    Unbiased scene graph generation from biased training,

    K. Tang, Y . Niu, J. Huang, J. Shi, and H. Zhang, “Unbiased scene graph generation from biased training,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 3716–3725

    2020

  17. [17]

    Scene graph generation by iterative message passing,

    D. Xu, Y . Zhu, C. B. Choy, and L. Fei-Fei, “Scene graph generation by iterative message passing,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2017, pp. 5410–5419

    2017

  18. [18]

    Unbiased scene graph generation in videos,

    S. Nag, K. Min, S. Tripathi, and A. K. Roy-Chowdhury, “Unbiased scene graph generation in videos,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 22 803–22 813

    2023

  19. [19]

    Oed: Towards one-stage end-to- end dynamic scene graph generation,

    G. Wang, Z. Li, Q. Chen, and Y . Liu, “Oed: Towards one-stage end-to- end dynamic scene graph generation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 27 938–27 947

    2024

  20. [20]

    Panoptic video scene graph generation,

    J. Yang, W. Peng, X. Li, Z. Guo, L. Chen, B. Li, Z. Ma, K. Zhou, W. Zhang, C. C. Loyet al., “Panoptic video scene graph generation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 18 675–18 685

    2023

  21. [21]

    Motion-aware contrastive learning for temporal panoptic scene graph generation,

    T. T. Nguyen, X. Wu, Y . Bin, C.-D. T. Nguyen, S.-K. Ng, and A. T. Luu, “Motion-aware contrastive learning for temporal panoptic scene graph generation,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 39, no. 6, 2025, pp. 6218–6226

    2025

  22. [22]

    Cyclo: Cyclic graph transformer approach to multi-object relationship modeling in aerial videos,

    T.-T. Nguyen, P. Nguyen, X. Li, J. Cothren, A. Yilmaz, and K. Luu, “Cyclo: Cyclic graph transformer approach to multi-object relationship modeling in aerial videos,”Advances in Neural Information Processing Systems, vol. 37, pp. 90 355–90 383, 2024

    2024

  23. [23]

    Rest: Holistic learning for end-to-end semantic segmentation 16 of whole-scene remote sensing imagery,

    W. Chen, L. Bruzzone, B. Dang, Y . Gao, Y . Deng, J.-G. Yu, L. Yuan, and Y . Li, “Rest: Holistic learning for end-to-end semantic segmentation 16 of whole-scene remote sensing imagery,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 48, no. 1, pp. 693–710, 2026

    2026

  24. [24]

    Full-scope vectorization of geographical elements from large- size remote sensing imagery,

    Y . Li, W. Li, B. Dang, Y . Wang, W. Chen, L. Wang, B. Yang, and Y . Zhang, “Full-scope vectorization of geographical elements from large- size remote sensing imagery,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 48, no. 6, pp. 6897–6911, 2026

    2026

  25. [25]

    Detecting and tracking small and dense moving objects in satellite videos: A benchmark,

    Q. Yin, Q. Hu, H. Liu, F. Zhang, Y . Wang, Z. Lin, W. An, and Y . Guo, “Detecting and tracking small and dense moving objects in satellite videos: A benchmark,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–18, 2021

    2021

  26. [26]

    Satsot: A benchmark dataset for satellite video single object tracking,

    M. Zhao, S. Li, S. Xuan, L. Kou, S. Gong, and Z. Zhou, “Satsot: A benchmark dataset for satellite video single object tracking,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–11, 2022

    2022

  27. [27]

    Deep learning-based object tracking in satellite videos: A comprehensive survey with a new dataset,

    Y . Li, L. Jiao, Z. Huang, X. Zhang, R. Zhang, X. Song, C. Tian, Z. Zhang, F. Liu, S. Yanget al., “Deep learning-based object tracking in satellite videos: A comprehensive survey with a new dataset,”IEEE Geoscience and Remote Sensing Magazine, vol. 10, no. 4, pp. 181–212, 2022

    2022

  28. [28]

    Multi-object tracking in satel- lite videos with graph-based multitask modeling,

    Q. He, X. Sun, Z. Yan, B. Li, and K. Fu, “Multi-object tracking in satel- lite videos with graph-based multitask modeling,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–13, 2022

    2022

  29. [29]

    Structural consistency- based multi-object detection for satellite video and benchmark with real scene simulation dataset,

    Y . Zhang, Y . Tang, F. Liu, Y . Chen, and T. Han, “Structural consistency- based multi-object detection for satellite video and benchmark with real scene simulation dataset,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 18, pp. 23 728–23 743, 2025

    2025

  30. [30]

    Visual relationship detection with language priors,

    C. Lu, R. Krishna, M. Bernstein, and L. Fei-Fei, “Visual relationship detection with language priors,” inProceedings of the European Con- ference on Computer Vision, 2016, pp. 852–869

    2016

  31. [31]

    Hi- erarchical memory learning for fine-grained scene graph generation,

    Y . Deng, Y . Li, Y . Zhang, X. Xiang, J. Wang, J. Chen, and J. Ma, “Hi- erarchical memory learning for fine-grained scene graph generation,” in Proceedings of the European Conference on Computer Vision. Springer, 2022, pp. 266–283

    2022

  32. [32]

    Image retrieval using scene graphs,

    J. Johnson, R. Krishna, M. Stark, L.-J. Li, D. Shamma, M. Bernstein, and L. Fei-Fei, “Image retrieval using scene graphs,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2015, pp. 3668–3678

    2015

  33. [33]

    Auto-encoding scene graphs for image captioning,

    X. Yang, K. Tang, H. Zhang, and J. Cai, “Auto-encoding scene graphs for image captioning,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 10 685–10 694

    2019

  34. [34]

    Relation-aware graph attention network for visual question answering,

    L. Li, Z. Gan, Y . Cheng, and J. Liu, “Relation-aware graph attention network for visual question answering,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 10 312–10 321

    2019

  35. [35]

    Recognition using visual phrases,

    M. A. Sadeghi and A. Farhadi, “Recognition using visual phrases,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2011, pp. 1745–1752

    2011

  36. [36]

    Visual genome: Connecting language and vision using crowdsourced dense image annotations,

    R. Krishna, Y . Zhu, O. Groth, J. Johnson, K. Hata, J. Kravitz, S. Chen, Y . Kalantidis, L.-J. Li, D. A. Shammaet al., “Visual genome: Connecting language and vision using crowdsourced dense image annotations,” International Journal of Computer Vision, vol. 123, pp. 32–73, 2017

    2017

  37. [37]

    Gqa: A new dataset for real-world visual reasoning and compositional question answering,

    D. A. Hudson and C. D. Manning, “Gqa: A new dataset for real-world visual reasoning and compositional question answering,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2019, pp. 6700–6709

    2019

  38. [38]

    The open images dataset v4: Unified image classification, object detection, and visual relationship detection at scale,

    A. Kuznetsova, H. Rom, N. Alldrin, J. Uijlings, I. Krasin, J. Pont-Tuset, S. Kamali, S. Popov, M. Malloci, A. Kolesnikovet al., “The open images dataset v4: Unified image classification, object detection, and visual relationship detection at scale,”International Journal of Computer Vision, vol. 128, no. 7, pp. 1956–1981, 2020

    1956

  39. [39]

    Panoptic scene graph generation,

    J. Yang, Y . Z. Ang, Z. Guo, K. Zhou, W. Zhang, and Z. Liu, “Panoptic scene graph generation,” inProceedings of the European Conference on Computer Vision, 2022, pp. 178–196

    2022

  40. [40]

    Visual relationship detection with deep structural ranking,

    K. Liang, Y . Guo, H. Chang, and X. Chen, “Visual relationship detection with deep structural ranking,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 32, no. 1, 2018, pp. 7098–7105

    2018

  41. [41]

    Stacked hybrid- attention and group collaborative learning for unbiased scene graph generation,

    X. Dong, T. Gan, X. Song, J. Wu, Y . Cheng, and L. Nie, “Stacked hybrid- attention and group collaborative learning for unbiased scene graph generation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 19 427–19 436

    2022

  42. [42]

    Detecting visual relationships with deep relational networks,

    B. Dai, Y . Zhang, and D. Lin, “Detecting visual relationships with deep relational networks,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2017, pp. 3076–3086

    2017

  43. [43]

    Neural motifs: Scene graph parsing with global context,

    R. Zellers, M. Yatskar, S. Thomson, and Y . Choi, “Neural motifs: Scene graph parsing with global context,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 5831–5840

    2018

  44. [44]

    Tensorize, factorize and regularize: Robust visual relationship learning,

    S. J. Hwang, S. N. Ravi, Z. Tao, H. J. Kim, M. D. Collins, and V . Singh, “Tensorize, factorize and regularize: Robust visual relationship learning,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 1014–1023

    2018

  45. [45]

    Visual relationship detection with internal and external linguistic knowledge distillation,

    R. Yu, A. Li, V . I. Morariu, and L. S. Davis, “Visual relationship detection with internal and external linguistic knowledge distillation,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2017, pp. 1974–1982

    2017

  46. [46]

    Bridging knowledge graphs to generate scene graphs,

    A. Zareian, S. Karaman, and S.-F. Chang, “Bridging knowledge graphs to generate scene graphs,” inProceedings of the European Conference on Computer Vision. Springer, 2020, pp. 606–623

    2020

  47. [47]

    Factorizable net: an efficient subgraph-based framework for scene graph generation,

    Y . Li, W. Ouyang, B. Zhou, J. Shi, C. Zhang, and X. Wang, “Factorizable net: an efficient subgraph-based framework for scene graph generation,” inProceedings of the European Conference on Computer Vision, 2018, pp. 335–351

    2018

  48. [48]

    Bipartite graph network with adaptive message passing for unbiased scene graph generation,

    R. Li, S. Zhang, B. Wan, and X. He, “Bipartite graph network with adaptive message passing for unbiased scene graph generation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 11 109–11 119

    2021

  49. [49]

    Unbiased scene graph generation via two-stage causal modeling,

    S. Sun, S. Zhi, Q. Liao, J. Heikkil ¨a, and L. Liu, “Unbiased scene graph generation via two-stage causal modeling,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 10, pp. 12 562– 12 580, 2023

    2023

  50. [50]

    Fine-grained scene graph generation via sample-level bias prediction,

    Y . Li, T. Wang, K. Wu, L. Wang, X. Guo, and W. Wang, “Fine-grained scene graph generation via sample-level bias prediction,” inProceedings of the European Conference on Computer Vision. Springer, 2024, pp. 18–35

    2024

  51. [51]

    A causal adjustment module for debiasing scene graph generation,

    L. Liu, S. Sun, S. Zhi, F. Shi, Z. Liu, J. Heikkil ¨a, and Y . Liu, “A causal adjustment module for debiasing scene graph generation,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 47, no. 5, pp. 4024–4043, 2025

    2025

  52. [52]

    Diffvsgg: Diffusion-driven online video scene graph generation,

    M. Chen, L. Li, W. Wang, and Y . Yang, “Diffvsgg: Diffusion-driven online video scene graph generation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025, pp. 29 161–29 172

    2025

  53. [53]

    End- to-end video scene graph generation with temporal propagation trans- former,

    Y . Zhang, Y . Pan, T. Yao, R. Huang, T. Mei, and C.-W. Chen, “End- to-end video scene graph generation with temporal propagation trans- former,”IEEE Transactions on Multimedia, vol. 26, pp. 1613–1625, 2023

    2023

  54. [54]

    Meta spatio-temporal debiasing for video scene graph generation,

    L. Xu, H. Qu, J. Kuen, J. Gu, and J. Liu, “Meta spatio-temporal debiasing for video scene graph generation,” inProceedings of the European Conference on Computer Vision. Springer, 2022, pp. 374– 390

    2022

  55. [55]

    Spatial–temporal knowledge- embedded transformer for video scene graph generation,

    T. Pu, T. Chen, H. Wu, Y . Lu, and L. Lin, “Spatial–temporal knowledge- embedded transformer for video scene graph generation,”IEEE Trans- actions on Image Processing, vol. 33, pp. 556–568, 2023

    2023

  56. [56]

    Td 2-net: Toward denoising and debiasing for video scene graph generation,

    X. Lin, C. Shi, Y . Zhan, Z. Yang, Y . Wu, and D. Tao, “Td 2-net: Toward denoising and debiasing for video scene graph generation,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 4, 2024, pp. 3495–3503

    2024

  57. [57]

    Semantic relation model and dataset for remote sensing scene understanding,

    P. Li, D. Zhang, A. Wulamu, X. Liu, and P. Chen, “Semantic relation model and dataset for remote sensing scene understanding,”ISPRS International Journal of Geo-Information, vol. 10, no. 7, p. 488, 2021

    2021

  58. [58]

    Message- passing-driven triplet representation for geo-object relational inference in hrsi,

    J. Chen, X. Zhou, Y . Zhang, G. Sun, M. Deng, and H. Li, “Message- passing-driven triplet representation for geo-object relational inference in hrsi,”IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1–5, 2021

    2021

  59. [59]

    Srsg and s2sg: A model and a dataset for scene graph generation of remote sensing images from segmentation results,

    Z. Lin, F. Zhu, Y . Kong, Q. Wang, and J. Wang, “Srsg and s2sg: A model and a dataset for scene graph generation of remote sensing images from segmentation results,”IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–11, 2022

    2022

  60. [60]

    Star: A first-ever dataset and a large-scale bench- mark for scene graph generation in large-size satellite imagery,

    Y . Li, L. Wang, T. Wang, X. Yang, J. Luo, Q. Wang, Y . Deng, W. Wang, X. Sun, H. Liet al., “Star: A first-ever dataset and a large-scale bench- mark for scene graph generation in large-size satellite imagery,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 47, no. 3, pp. 1832–1849, 2025

    2025

  61. [61]

    Recon1m: A large-scale benchmark dataset for relation comprehension in remote sensing imagery,

    Q. Yan, C. Deng, C. Liu, Z. Hou, X. Liu, Y . Jiang, W. Lu, F. Yao, X. Liu, L. Haoet al., “Recon1m: A large-scale benchmark dataset for relation comprehension in remote sensing imagery,”IEEE Transactions on Geoscience and Remote Sensing, vol. 63, pp. 1–22, 2025

    2025

  62. [62]

    Scene-specific multi-prototype network for remote sensing scene graph generation,

    Z. Hou, C. Deng, Q. Yan, X. Liu, T. Ling, W. Lu, H. Yu, Y . Hou, and X. Sun, “Scene-specific multi-prototype network for remote sensing scene graph generation,”IEEE Transactions on Geoscience and Remote Sensing, vol. 63, pp. 1–21, 2025

    2025

  63. [63]

    Hierarchical prototype learning via aggregation-decomposition for fine-grained geospatial scene graph 17 generation,

    L. Wang, T. Wang, J. Luo, and Y . Li, “Hierarchical prototype learning via aggregation-decomposition for fine-grained geospatial scene graph 17 generation,”IEEE Transactions on Geoscience and Remote Sensing, vol. 63, pp. 1–13, 2025

    2025

  64. [64]

    Bias-aware learning for unbiased scene graph generation in remote sensing imagery,

    T. Wang, L. Wang, J. Luo, K. Wu, and Y . Li, “Bias-aware learning for unbiased scene graph generation in remote sensing imagery,”ISPRS Journal of Photogrammetry and Remote Sensing, vol. 231, pp. 473–486, 2026

    2026

  65. [65]

    Detection of event of interest for satellite video understanding,

    Y . Gu, T. Wang, X. Jin, and G. Gao, “Detection of event of interest for satellite video understanding,”IEEE Transactions on Geoscience and Remote Sensing, vol. 58, no. 11, pp. 7860–7871, 2020

    2020

  66. [66]

    Recent advances in intelligent processing of satellite video: Challenges, methods, and applications,

    S. Li, X. Sun, Y . Gu, Y . Lv, M. Zhao, Z. Zhou, W. Guo, Y . Sun, H. Wang, and J. Yang, “Recent advances in intelligent processing of satellite video: Challenges, methods, and applications,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 16, pp. 6776– 6798, 2023

    2023

  67. [67]

    Dual-path self-supervised spa- tiotemporal learning for satellite video scene understanding,

    Y . Wang, J. Li, G. Gao, and Y . Gu, “Dual-path self-supervised spa- tiotemporal learning for satellite video scene understanding,”IEEE Transactions on Geoscience and Remote Sensing, vol. 64, pp. 1–11, 2026

    2026

  68. [68]

    Target-aware transformer for satellite video object tracking,

    P. Lai, M. Zhang, G. Cheng, S. Li, X. Huang, and J. Han, “Target-aware transformer for satellite video object tracking,”IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–10, 2024

    2024

  69. [69]

    An empirical evaluation of generic convolutional and recurrent networks for sequence modeling,

    S. Bai, J. Z. Kolter, and V . Koltun, “An empirical evaluation of generic convolutional and recurrent networks for sequence modeling,”arXiv preprint arXiv:1803.01271, 2018

    Pith/arXiv arXiv 2018

  70. [70]

    Unbiased video scene graph generation via visual and semantic dual debiasing,

    Y . Li, Z. Li, H. Chen, and L. Xu, “Unbiased video scene graph generation via visual and semantic dual debiasing,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025, pp. 19 047–19 056

    2025

  71. [71]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” inInternational Conference on Learning Representations, 2015

    2015

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.