Pith. sign in

REVIEW 4 major objections 6 minor 63 references

ScanFocus claims that decoupling spatio-temporal video grounding into a coarse global scan and a dense local boundary-focus stage, driven by the Semantic-Guided Temporal Aggregator, recovers the high-frequency boundary cues that low-rate do

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 ↗

A coarse-to-fine video-grounding framework improves temporal boundary accuracy by densely re-examining frames around coarse start/end predictions, achieving SOTA on HC-STVGv1/v2 and VidSTG.

T0 review reviewed 2026-08-02 challenge →

load-bearing objection A solid, well-ablated coarse-to-fine STVG paper with a real but unmeasured assumption about how close the coarse stage needs to be; worth a serious referee. the 4 major comments →

arxiv 2607.13421 v1 pith:LNIWPIZJ submitted 2026-07-15 cs.CV cs.AI

ScanFocus: A Coarse-to-Fine Framework for Spatio-Temporal Video Grounding

Kai Chen , Ming Dai , Wenxuan Cheng , Wankou Yang This is my paper
classification cs.CV cs.AI
keywords spatio-temporal video groundingcoarse-to-fineboundary refinementtemporal localizationsemantic-guided temporal aggregatordense samplingmultimodal fusionvideo grounding
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 argues that the main obstacle to accurate spatio-temporal video grounding—locating the trajectory of an object described in a sentence, frame by frame—is not spatial detection but temporal boundary ambiguity, caused by the low-rate temporal downsampling that most existing methods use to keep global modeling affordable. ScanFocus tries to prove that decoupling the task into a coarse global scan at a low frame rate and a fine local refinement that densely samples the predicted boundaries recovers the high-frequency cues that downsampling suppresses. At the core is the Semantic-Guided Temporal Aggregator (SGTA), which explicitly models short-term inter-frame dependencies inside each dense boundary window, guided by the referred object's appearance and the language query. The paper reports consistent state-of-the-art results on HC-STVGv1/v2 and VidSTG, with the largest gains on strict metrics like vIoU@0.3 and vIoU@0.5, and an oracle experiment indicating that temporal localization, not spatial localization, is the current bottleneck.

Core claim

ScanFocus claims that a coarse-to-fine decomposition of spatio-temporal video grounding—first a sparse global scan that produces a rough temporal interval, then a local boundary-focus stage that densely samples frames around that interval and explicitly models inter-frame dependencies—recovers the high-frequency boundary cues lost to standard low-rate temporal subsampling, and that this recovery is what drives precise start/end timestamp regression. The central evidence is a set of benchmark results: on HC-STVGv1, the method reaches m_tIoU 55.5, m_vIoU 41.8, vIoU@0.3 67.5, and vIoU@0.5 42.2, surpassing the previous best (TA-STVG) by 4.4 and 5.4 points on vIoU@0.3 and vIoU@0.5 respectively; c

What carries the argument

The key machinery is the local boundary-focus stage built around the Semantic-Guided Temporal Aggregator (SGTA). Given a coarse predicted interval, the method densely samples Nw=8 frames at twice the coarse frame rate in two windows centered on the predicted start and end. SGTA then: (1) uses the coarse bounding-box sequence, temporally interpolated to the window length, as a spatial prior for RoI pooling on the appearance features; (2) modulates the motion features by the object-aligned appearance features (Hadamard product) and injects the text features as a bias; and (3) flattens the resulting semantic-guided motion tokens into a sequence and passes them through stacked temporal self-atte

Load-bearing premise

The refinement stage can only fix boundaries if the true start and end frames fall inside the dense sampling windows centered on the coarse predictions; if the coarse interval is off by more than about half a window, the refinement never sees the true boundary, and the paper does not measure how often that happens.

What would settle it

Compute the fraction of ground-truth start and end timestamps that lie within the Nw=8 dense windows (sampled at twice the coarse rate) around the model's coarse predictions. If that recall is significantly below 100%—say, under 90%—then the refinement stage is provably unable to recover those missed boundaries, and the gain attributed to SGTA would overstate the recoverable improvement. A complementary experiment: feed the refine stage with coarse intervals artificially perturbed by increasing offsets (0, 2, 4, 6, 8 frames) and measure how quickly precision degrades; the degradation curve rev

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

If this is right

  • If the decoupling claim holds, temporal boundary precision in spatio-temporal video grounding can be improved without paying the full cost of dense global processing, since the dense computation is confined to small windows around coarse boundaries.
  • The oracle experiment implies that the spatial grounding branch is already near-saturated; further gains on these benchmarks are most likely to come from better temporal boundary modeling, not better object detection.
  • The SGTA design—semantic guidance via RoI pooling on interpolated coarse boxes plus temporal self-attention on flattened tokens—offers a reusable template for other video tasks that need fine temporal boundaries, such as action localization or moment retrieval.
  • Replacing the heavy MDETR-style tri-modal fusion with a general pre-trained vision-language encoder plus a lightweight deformable fusion reduces computational cost while improving accuracy, suggesting that decoupling perception and fusion can be beneficial beyond this task.
  • The reported gains are largest on strict temporal-spatial overlap thresholds (vIoU@0.3 and vIoU@0.5), indicating that the method specifically improves boundary sharpness rather than just coarse region agreement.

Where Pith is reading between the lines

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

  • The paper's fixed window size (Nw=8 at twice the coarse rate) implies a hard reach: any true boundary more than half a window away from the coarse prediction is invisible to the refinement stage. A natural testable consequence is that the method's gains should concentrate on clips where the coarse prediction is already within a few frames of the true boundary; if that recall is low, the reported m
  • The same coarse-to-fine recipe could be applied in an iterative or cascaded manner (refine the refined boundaries again), or with adaptive window sizes based on coarse confidence, which the paper does not explore.
  • Because the refinement stage reuses the coarse features and only adds lightweight modules, the approach suggests that the bottleneck in many video-language tasks may be temporal granularity rather than model capacity; a cheaper fix for existing systems might be to add a similar boundary-refinement head rather than retrain the whole model.
  • If the oracle gap is as large as reported (vIoU@0.3 jumps to 94% with ground-truth timestamps), then even a perfect temporal-refinement module would still leave substantial room before spatial grounding becomes the limiting factor; the paper's framing implies that future work should focus on boundary regression accuracy, perhaps with explicit uncertainty estimation.
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

4 major / 6 minor

Summary. ScanFocus proposes a two-stage coarse-to-fine framework for spatio-temporal video grounding (STVG). A global stage uniformly samples T_c=64 frames at a low rate, fuses BEiT-3 appearance/text features with VideoMAE motion features via a lightweight deformable semantic-motion fusion encoder, and generates coarse spatial tubes and a coarse temporal interval with dual DETR-style decoders. A local refinement stage then densely samples N_w=8 frames at twice the coarse rate around each predicted boundary, applies the Semantic-Guided Temporal Aggregator (SGTA) to model short-term inter-frame dependencies, and regresses refined start/end timestamps. The paper reports state-of-the-art numbers on HC-STVGv1, HC-STVGv2, and VidSTG, an oracle experiment attributing most remaining error to temporal localization, and ablations over components, window size, SGTA design, and fusion mechanism.

Significance. If the reported results are reproducible, ScanFocus is a solid contribution to STVG: it targets a real limitation of low-frame-rate global modeling, provides a clean architecture that decouples global context from local boundary precision, and reports consistent gains across three benchmarks. The oracle analysis in Fig. 2 is a valuable diagnostic, and the component ablations in Tables 4–7 generally support the design choices; the efficiency gain in Table 7 is also a concrete plus. The main risk is that the refinement stage's central assumption—that coarse boundaries fall within ±1.75 coarse sampling intervals—is not validated. Without a hit-rate analysis, the mechanism claimed to recover high-frequency cues is not actually demonstrated.

major comments (4)
  1. [Sec. 3.4 (Dense Boundary Sampling)] The refinement windows sample N_w=8 frames at f_r=2·f_c centered on the predicted coarse boundaries. The largest boundary offset that can possibly be recovered is therefore (N_w−1)/(2·f_r)=7/(4·f_c)≈1.75 coarse frame intervals. The paper never reports the distribution of coarse-boundary errors, the fraction of GT boundaries that fall inside the sampled windows, or whether refine-stage training windows are centered on predicted or GT coarse boundaries. Without this, the 2.8-point tIoU improvement in Table 4 (50.9→53.7) and the SOTA claim remain consistent with refinement helping only samples with already-accurate coarse predictions; GT-centered training would additionally create a train/test mismatch. Please report coarse-interval recall / window hit-rate and the exact training window centering.
  2. [Sec. 4.3, Tables 4–6] Tables 4–6 report single runs with no error bars, standard deviations, or significance tests. Some of the claims rest on small differences: Full SGTA vs w/o SG is 0.6 tIoU and vs w/o TA is 0.9 tIoU in Table 6. Moreover, ablations are run at 224×224, while the main comparisons in Tables 1–3 use 384×384 and a different backbone than prior methods, so the apparent SOTA margin may partly reflect the choice of BEiT-3/VideoMAE rather than the coarse-to-fine design. Please report multi-seed statistics and a same-backbone, same-resolution baseline (e.g., the coarse global stage without refinement) in the main results.
  3. [Sec. 4.1 (Implementation)] Sec. 4.1 states that detailed network configurations and loss weights are provided in the Supplementary Material, but no supplementary document is included with arXiv v1, and the paper also refers to it for the other datasets' oracle experiments (Sec. 4.1, Fig. 2 caption). Without specifying the number of encoder/decoder layers, attention heads, loss weights λ_box/λ_tmp/λ_ref/λ_act, learning rates, and training iterations, the experiments are not reproducible. The refine-stage training protocol is especially important because it determines whether the model sees predicted or GT windows during training.
  4. [Sec. 3.3 (Feature Extraction)] Motion features are defined as F_m ∈ R^{T×N_m×C_m} with T the original video length, while the coarse input consists of only T_c uniformly sampled frames. If VideoMAE is applied to the full T frames, the claimed computational savings from sparse sampling are negated; if it is applied to the sampled frames, the notation is inconsistent and the alignment between F_m and F_a (which has T_c frames) needs to be stated precisely. The same issue affects the complexity analysis in Table 7, which reports GFLOPs but not the temporal resolution used for motion extraction.
minor comments (6)
  1. [Sec. 3.4] First sentence: 'motion details that suppressed by' should be 'that are suppressed by'.
  2. [Eq. (5)] The projections φ_a and φ_t are not dimensionally specified; please give the output dimensions of φ_a(F_ps^a) and φ_t(F_rs^t), or state that all features are projected to a common channel C.
  3. [Fig. 3] The dense-boundary-sampling region contains many unreadable small labels (e.g., 'cc_132509', '1424', the list of numbers); please enlarge fonts and remove irrelevant tokens.
  4. [Sec. 4.3] 'For brevity, tiou and viou are used to represent m_tIoU and m_vIoU'—define these abbreviations at first use and use consistent capitalization.
  5. [Sec. 4.1] Inputs are resized to 384×384 for main results but ablations use 224×224. Please clarify whether the GFLOPs in Tables 5 and 7 correspond to the 224×224 or 384×384 setting.
  6. [References] Reference [46] has a typo ('Protege'); reference [55] contains a stray '(7)' in the arXiv ID. Please verify the official bibliographic details.

Circularity Check

0 steps flagged

No significant circularity: benchmark claims are empirical; refinement stage is not definitionally tied to coarse outputs.

full rationale

ScanFocus's central claim is an empirical SOTA result on HC-STVGv1/v2 and VidSTG, not a derivation. The coarse-to-fine refinement uses the coarse interval only to center dense windows (Sec. 3.4); the refine decoder independently regresses start/end probabilities (Eq. 8) and can output timestamps different from the coarse ones, so the prediction is not forced by construction. The only self-citation by overlapping authors is ref. [12] (GC3VG), cited in a list of visual-grounding works (Sec. 2.2) and not used to justify the framework, its equations, or the choice of coarse-to-fine structure; it is therefore not load-bearing. Hyperparameter selection of Nw=8 on HC-STVGv1 (Table 5) is benchmark tuning, not a fitted-parameter-called-prediction. The oracle experiment (Fig. 2) uses GT timestamps as an analysis of the spatial/temporal bottleneck, not as a claimed prediction. The unvalidated recoverable-range assumption for the dense windows is a correctness risk, not circularity.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

The central claim rests on standard pretrained encoders (BEiT-3, VideoMAE), a few hand-set or benchmark-tuned hyperparameters (Nw, kr, Tc, loss weights), and the unquantified assumption that coarse boundary windows contain the true boundaries. No novel physical or conceptual entities are introduced; SGTA is an architectural module, not an invented entity.

free parameters (4)
  • Nw (refine window size) = 8
    Selected by sweeping Nw in {4,6,8,10,12} on HC-STVGv1 (Table 5) and then used as the default for all datasets; this is tuning on the benchmark.
  • kr (refine density factor) = 2
    Set in Sec. 4.1 Implementation; controls how many dense frames are sampled around boundaries, with no reported sensitivity analysis.
  • Tc (coarse frame count) = 64
    Number of uniformly sampled coarse frames, chosen without a reported ablation.
  • Loss weights lambda_box, lambda_tmp, lambda_ref, lambda_act = not given in paper
    Balance the losses in Eq. (7)-(8); values are referred to the missing Supplementary Material.
axioms (5)
  • domain assumption BEiT-3's unified vision-language pretraining provides features already aligned across visual and textual modalities (Sec. 3.3).
    The method relies on this pretrained alignment to replace the MDETR-style tri-modal fusion and does not validate the choice against a same-backbone baseline.
  • domain assumption Frozen VideoMAE motion features adequately encode motion for STVG (Sec. 3.3).
    Motion features are extracted by a frozen VideoMAE and never fine-tuned; if they lack boundary-relevant motion cues, SGTA cannot recover them.
  • domain assumption Coarse predicted boundary windows cover the true boundaries often enough (Sec. 3.4).
    Dense sampling happens only around coarse boundaries; the paper provides no coverage statistics or coarse-interval recall analysis.
  • domain assumption Deformable attention in Eq. (2) with learnable offsets preserves the information needed for temporal boundary localization (Sec. 3.3).
    Sparse deformable sampling could miss salient boundary tokens; no analysis of sampling-point coverage is provided.
  • domain assumption Benchmark splits and reported baselines are directly comparable (Sec. 4.1).
    Results are compared with numbers taken from other papers; no same-codebase reruns, seed variance, or statistical tests are reported.

reviewed 2026-08-02 · how reviews work

0 comments
Cite this review

Pith. "Pith review of ScanFocus: A Coarse-to-Fine Framework for Spatio-Temporal Video Grounding." pith.science (2026). https://pith.science/paper/LNIWPIZJ

@misc{pith2026260713421,
  author       = {Pith},
  title        = {Pith review of: ScanFocus: A Coarse-to-Fine Framework for Spatio-Temporal Video Grounding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LNIWPIZJ}},
  note         = {Machine review of arXiv:2607.13421}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Spatio-Temporal Video Grounding (STVG) aims to retrieve the visual trajectory of a specific object from a video stream as described by a natural language expression. However, most advanced methods struggle to balance global context modeling with precise boundary localization. Due to the prohibitive computational costs of processing long videos, these approaches typically resort to low-rate temporal downsampling and implicit motion modeling. This inevitably suppresses high-frequency boundary cues and neglects the explicit inter-frame dependencies required for precise boundary delineation. To address these limitations, we present \textbf{ScanFocus}, a novel coarse-to-fine framework that decouples the STVG task into a global spatio-temporal scan and a local boundary focus. Specifically, we utilize a unified vision-language fusion encoder combined with a lightweight Deformable Semantic-Motion Fusion module to efficiently align multimodal features and generate coarse proposals. To recover the suppressed fine-grained details, we introduce the Semantic-Guided Temporal Aggregator (SGTA) in the refinement stage. By densely sampling around coarse boundaries, SGTA explicitly models short-term temporal interactions under semantic guidance, capturing rapid motion changes for precise timestamp regression. Extensive experiments on three widely used benchmarks demonstrate the performance superiority of our proposed method over previous approaches. Code will be released at https://github.com/TenMinutes209/ScanFocus.

Figures

Figures reproduced from arXiv: 2607.13421 by Kai Chen, Ming Dai, Wankou Yang, Wenxuan Cheng.

Figure 1
Figure 1. Figure 1: Comparison of temporal boundary localization paradigms. (a) Existing Transformer-based methods often produce ambiguous boundaries due to the suppres￾sion of high-frequency temporal cues caused by global downsampling. (b) Our proposed method adopts a coarse-to-fine framework that first generates a coarse interval at a low frame rate, followed by boundary dense sampling to recover fine-grained details for pr… view at source ↗
Figure 2
Figure 2. Figure 2: Oracle Experiment on HCSTVG-v1. Using GT timestamps instead of predicted intervals substantially improves performance, identifying tempo￾ral localization as the main bottleneck. To empirically validate this tem￾poral localization bottleneck, we conduct an oracle analysis on the HC￾STVGv1 dataset (for other datasets, please kindly refer to Supplementary Material). As illustrated in [PITH_FULL_IMAGE:figures… view at source ↗
Figure 3
Figure 3. Figure 3: Overview architecture of our proposed ScanFocus. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of multi-modal fusion mechanisms. (a) [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Detailed architecture of SGTA module. Semantic-Guided Temporal Ag￾gregator. While the shared encoder effectively extracts frame-wise repre￾sentations, precise boundary delin￾eation inherently requires capturing fine-grained temporal evolution, which relies on explicit inter-frame reasoning. To bridge this gap, we introduce SGTA. Leveraging the reduced temporal foot￾print of the local windows (Nw), SGTA per… view at source ↗
Figure 6
Figure 6. Figure 6: Qualitative results of ScanFocus. The coarse stage often generates ambigu￾ous boundaries. Our refine stage successfully corrects these initial proposals to align precisely with the Ground Truth. Fund for Advanced Ocean Institute of Southeast University (Major Program MP202404). References 1. Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., … 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

63 extracted references · 13 linked inside Pith

  1. [1]

    arXiv preprint arXiv:2303.08774 (2023)

    Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al.: Gpt-4 technical report. arXiv preprint arXiv:2303.08774 (2023)

    Pith/arXiv arXiv 2023

  2. [2]

    In: ICCV

    Anne Hendricks, L., Wang, O., Shechtman, E., Sivic, J., Darrell, T., Russell, B.: Lo- calizing moments in video with natural language. In: ICCV. pp. 5803–5812 (2017)

    2017

  3. [3]

    In: ICCV

    Antol, S., Agrawal, A., Lu, J., Mitchell, M., Batra, D., Zitnick, C.L., Parikh, D.: Vqa: Visual question answering. In: ICCV. pp. 2425–2433 (2015)

    2015

  4. [4]

    arXiv preprint arXiv:2511.21631 (2025)

    Bai, S., Cai, Y., Chen, R., Chen, K., Chen, X., Cheng, Z., Deng, L., Ding, W., Gao, C., Ge, C., et al.: Qwen3-vl technical report. arXiv preprint arXiv:2511.21631 (2025)

    Pith/arXiv arXiv 2025

  5. [5]

    arXiv preprint arXiv:2106.08254 (2021)

    Bao, H., Dong, L., Piao, S., Wei, F.: Beit: Bert pre-training of image transformers. arXiv preprint arXiv:2106.08254 (2021)

    Pith/arXiv arXiv 2021

  6. [6]

    In: ICCV

    Barrios, W., Soldan, M., Ceballos-Arroyo, A.M., Heilbron, F.C., Ghanem, B.: Lo- calizing moments in long video via multimodal guidance. In: ICCV. pp. 13667– 13678 (2023)

    2023

  7. [7]

    In: ECCV

    Cao, M., Yang, T., Weng, J., Zhang, C., Wang, J., Zou, Y.: Locvtp: Video-text pre-training for temporal localization. In: ECCV. pp. 38–56. Springer (2022)

    2022

  8. [8]

    In: WACV

    Cao, Z., Zhang, B., Du, H., Yu, X., Li, X., Wang, S.: Flashvtg: Feature layering and adaptive score handling network for video temporal grounding. In: WACV. pp. 9226–9236. IEEE (2025)

    2025

  9. [9]

    In: ECCV

    Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., Zagoruyko, S.: End- to-end object detection with transformers. In: ECCV. pp. 213–229. Springer (2020)

    2020

  10. [10]

    Chen et al

    Chen, W., Chen, L., Wu, Y.: An efficient and effective transformer decoder-based frameworkformulti-taskvisualgrounding.In:ECCV.pp.125–141.Springer(2024) 16 K. Chen et al

    2024

  11. [11]

    NeurIPS34, 28442–28453 (2021)

    Chen, Y.W., Tsai, Y.H., Yang, M.H.: End-to-end multi-modal video temporal grounding. NeurIPS34, 28442–28453 (2021)

    2021

  12. [12]

    TCSVT (2025)

    Dai,M.,Chen,K.,Cheng,W.,Zhuang,J.,Feng,Z.,Zhu,P.,Yang,W.:Gc3vg:Gen- eralized multi-task visual grounding with coarse-to-fine consistency constraints. TCSVT (2025)

    2025

  13. [13]

    TPAMI (2025)

    Dai, M., Cheng, W., Liu, J.J., Yang, L., Feng, Z., Yang, W., Wang, J.: Improving generalized visual grounding with instance-aware joint learning. TPAMI (2025)

    2025

  14. [14]

    In: ICCV

    Dai, M., Cheng, W., Liu, J.j., Yang, S., Cai, W., Sun, Y., Yang, W.: Deris: Decou- pling perception and cognition for enhanced referring image segmentation through loopback synergy. In: ICCV. pp. 19936–19946 (2025)

    2025

  15. [15]

    NeurIPS37, 121670–121698 (2024)

    Dai, M., Yang, L., Xu, Y., Feng, Z., Yang, W.: Simvg: A simple framework for visual grounding with decoupled multi-modal fusion. NeurIPS37, 121670–121698 (2024)

    2024

  16. [16]

    arXiv preprint arXiv:2510.09274 (2025)

    Dai, M., Yang, S., Duan, B., Yang, W., Wang, J.: Momentseg: Moment-centric sampling for enhanced video pixel understanding. arXiv preprint arXiv:2510.09274 (2025)

    arXiv 2025

  17. [17]

    In: ICCV

    Gao, J., Sun, C., Yang, Z., Nevatia, R.: Tall: Temporal activity localization via language query. In: ICCV. pp. 5267–5275 (2017)

    2017

  18. [18]

    In: CVPR

    Gu, X., Chen, G., Wang, Y., Zhang, L., Luo, T., Wen, L.: Text with knowledge graph augmented transformer for video captioning. In: CVPR. pp. 18941–18951 (2023)

    2023

  19. [19]

    In: CVPR

    Gu, X., Fan, H., Huang, Y., Luo, T., Zhang, L.: Context-guided spatio-temporal video grounding. In: CVPR. pp. 18330–18339 (2024)

    2024

  20. [20]

    arXiv preprint arXiv:2502.11168 (2025)

    Gu, X., Shen, Y., Luo, C., Luo, T., Huang, Y., Lin, Y., Fan, H., Zhang, L.: Knowing your target: Target-aware transformer makes better spatio-temporal video ground- ing. arXiv preprint arXiv:2502.11168 (2025)

    Pith/arXiv arXiv 2025

  21. [21]

    arXiv preprint arXiv:2511.21375 (2025)

    Gu, X., Zhang, H., Fan, Q., Niu, J., Zhang, Z., Zhang, L., Chen, G., Chen, F., Wen, L., Zhu, S.: Thinking with bounding boxes: Enhancing spatio-temporal video grounding via reinforcement fine-tuning. arXiv preprint arXiv:2511.21375 (2025)

    arXiv 2025

  22. [22]

    arXiv preprint arXiv:2501.12948 (2025)

    Guo, D., Yang, D., Zhang, H., Song, J., Wang, P., Zhu, Q., Xu, R., Zhang, R., Ma, S., Bi, X., et al.: Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948 (2025)

    Pith/arXiv arXiv 2025

  23. [23]

    5-vl technical report

    Guo, D., Wu, F., Zhu, F., Leng, F., Shi, G., Chen, H., Fan, H., Wang, J., Jiang, J., Wang, J., et al.: Seed1. 5-vl technical report. arXiv preprint arXiv:2505.07062 (2025)

    Pith/arXiv arXiv 2025

  24. [24]

    In: ECCV

    Hao, J., Sun, H., Ren, P., Wang, J., Qi, Q., Liao, J.: Can shuffling video benefit temporal bias problem: A novel training framework for temporal grounding. In: ECCV. pp. 130–147. Springer (2022)

    2022

  25. [25]

    In: CVPR

    Jiang, H., Misra, I., Rohrbach, M., Learned-Miller, E., Chen, X.: In defense of grid features for visual question answering. In: CVPR. pp. 10267–10276 (2020)

    2020

  26. [26]

    NeurIPS35, 29192–29204 (2022)

    Jin, Y., Yuan, Z., Mu, Y., et al.: Embracing consistency: A one-stage approach for spatio-temporal video grounding. NeurIPS35, 29192–29204 (2022)

    2022

  27. [27]

    In: ICCV

    Kamath, A., Singh, M., LeCun, Y., Synnaeve, G., Misra, I., Carion, N.: Mdetr- modulated detection for end-to-end multi-modal understanding. In: ICCV. pp. 1780–1790 (2021)

    2021

  28. [28]

    In: ECCV

    Kang, W., Liu, G., Shah, M., Yan, Y.: Segvg: Transferring object bounding box to segmentation for visual grounding. In: ECCV. pp. 57–75. Springer (2024)

    2024

  29. [29]

    NeurIPS34, 11846–11858 (2021) ScanFocus: A Coarse-to-Fine Framework for STVG 17

    Lei, J., Berg, T.L., Bansal, M.: Detecting moments and highlights in videos via natural language queries. NeurIPS34, 11846–11858 (2021) ScanFocus: A Coarse-to-Fine Framework for STVG 17

    2021

  30. [30]

    In: CVPR

    Li, H., Chen, J., Wei, Z., Huang, S., Hui, T., Gao, J., Wei, X., Liu, S.: Llava-st: A multimodal large language model for fine-grained spatial-temporal understanding. In: CVPR. pp. 8592–8603 (2025)

    2025

  31. [31]

    IEEE transactions on medical imaging43(1), 96–107 (2023)

    Li, Z., Li, Y., Li, Q., Wang, P., Guo, D., Lu, L., Jin, D., Zhang, Y., Hong, Q.: Lvit: language meets vision transformer in medical image segmentation. IEEE transactions on medical imaging43(1), 96–107 (2023)

    2023

  32. [32]

    In: ICCV

    Lin, K.Q., Zhang, P., Chen, J., Pramanick, S., Gao, D., Wang, A.J., Yan, R., Shou, M.Z.: Univtg: Towards unified video-language temporal grounding. In: ICCV. pp. 2794–2804 (2023)

    2023

  33. [33]

    In: CVPR

    Lin, Z., Tan, C., Hu, J.F., Jin, Z., Ye, T., Zheng, W.S.: Collaborative static and dynamic vision-language streams for spatio-temporal video grounding. In: CVPR. pp. 23100–23109 (2023)

    2023

  34. [34]

    In: ECCV

    Liu, Y., He, J., Li, W., Kim, J., Wei, D., Pfister, H., Chen, C.W.: r 2-tuning: Efficient image-to-video transfer learning for video temporal grounding. In: ECCV. pp. 421–438. Springer (2024)

    2024

  35. [35]

    arXiv preprint arXiv:2311.08835 (2023)

    Moon, W., Hyun, S., Lee, S., Heo, J.P.: Correlation-guided query-dependency cal- ibration for video temporal grounding. arXiv preprint arXiv:2311.08835 (2023)

    Pith/arXiv arXiv 2023

  36. [36]

    In: CVPR

    Moon, W., Hyun, S., Park, S., Park, D., Heo, J.P.: Query-dependent video repre- sentation for moment retrieval and highlight detection. In: CVPR. pp. 23023–23033 (2023)

    2023

  37. [37]

    In: CVPR

    Mun, J., Cho, M., Han, B.: Local-global video-text interactions for temporal grounding. In: CVPR. pp. 10810–10819 (2020)

    2020

  38. [38]

    In: CVPR

    Ren, Z., Huang, Z., Wei, Y., Zhao, Y., Fu, D., Feng, J., Jin, X.: Pixellm: Pixel reasoning with large multimodal model. In: CVPR. pp. 26374–26383 (2024)

    2024

  39. [39]

    In: ICCV

    Shen,Y.,Gu,X.,Xu,K.,Fan,H.,Wen,L.,Zhang,L.:Accurateandfastcompressed video captioning. In: ICCV. pp. 15558–15567 (2023)

    2023

  40. [40]

    In: ICCV

    Su, R., Yu, Q., Xu, D.: Stvgbert: A visual-linguistic transformer based framework for spatio-temporal video grounding. In: ICCV. pp. 1533–1542 (2021)

    2021

  41. [41]

    arXiv preprint arXiv:2106.10634 (2021)

    Tan, C., Lin, Z., Hu, J.F., Li, X., Zheng, W.S.: Augmented 2d-tan: A two- stage approach for human-centric spatio-temporal video grounding. arXiv preprint arXiv:2106.10634 (2021)

    Pith/arXiv arXiv 2021

  42. [42]

    TCSVT32(12), 8238–8249 (2021)

    Tang, Z., Liao, Y., Liu, S., Li, G., Jin, X., Jiang, H., Yu, Q., Xu, D.: Human- centric spatio-temporal video grounding with visual transformers. TCSVT32(12), 8238–8249 (2021)

    2021

  43. [43]

    NeurIPS35, 10078–10093 (2022)

    Tong, Z., Song, Y., Wang, J., Wang, L.: Videomae: Masked autoencoders are data- efficient learners for self-supervised video pre-training. NeurIPS35, 10078–10093 (2022)

    2022

  44. [44]

    NeurIPS30(2017)

    Vaswani,A.,Shazeer,N.,Parmar,N.,Uszkoreit,J.,Jones,L.,Gomez,A.N.,Kaiser, Ł., Polosukhin, I.: Attention is all you need. NeurIPS30(2017)

    2017

  45. [45]

    arXiv preprint arXiv:2503.13983 (2025)

    Wang,J.,Zhang,Z.,Liu,Z.,Li,Y.,Ge,J.,Xie,H.,Zhang,Y.:Spacevllm:Endowing multimodal large language model with spatio-temporal video grounding capability. arXiv preprint arXiv:2503.13983 (2025)

    Pith/arXiv arXiv 2025

  46. [46]

    In: CVPR

    Wang, L., Mittal, G., Sajeev, S., Yu, Y., Hall, M., Boddeti, V.N., Chen, M.: Protege: Untrimmed pretraining for video temporal grounding by video tempo- ral grounding. In: CVPR. pp. 6575–6585 (2023)

    2023

  47. [47]

    In: ACMMM

    Wang, W., Liu, J., Su, Y., Nie, W.: Efficient spatio-temporal video grounding with semantic-guided feature decomposition. In: ACMMM. pp. 4867–4876 (2023)

    2023

  48. [48]

    In: CVPR

    Wang, W., Bao, H., Dong, L., Bjorck, J., Peng, Z., Liu, Q., Aggarwal, K., Mo- hammed, O.K., Singhal, S., Som, S., et al.: Image as a foreign language: Beit pre- training for vision and vision-language tasks. In: CVPR. pp. 19175–19186 (2023) 18 K. Chen et al

    2023

  49. [49]

    In: AAAI

    Wang, Z., Wang, L., Wu, T., Li, T., Wu, G.: Negative sample matters: A renais- sance of metric learning for temporal grounding. In: AAAI. vol. 36, pp. 2613–2623 (2022)

    2022

  50. [50]

    In: CVPR

    Wasim, S.T., Naseer, M., Khan, S., Yang, M.H., Khan, F.S.: Videogrounding-dino: Towards open-vocabulary spatio-temporal video grounding. In: CVPR. pp. 18909– 18918 (2024)

    2024

  51. [51]

    In: ECCV

    Weng, Y., Han, M., He, H., Chang, X., Zhuang, B.: Longvlm: Efficient long video understanding via large language models. In: ECCV. pp. 453–470. Springer (2024)

    2024

  52. [52]

    In: AAAI

    Xu, H., He, K., Plummer, B.A., Sigal, L., Sclaroff, S., Saenko, K.: Multilevel lan- guage and vision integration for text-to-clip retrieval. In: AAAI. vol. 33, pp. 9062– 9069 (2019)

    2019

  53. [53]

    In: CVPR

    Yang, A., Miech, A., Sivic, J., Laptev, I., Schmid, C.: Tubedetr: Spatio-temporal video grounding with transformers. In: CVPR. pp. 16442–16453 (2022)

    2022

  54. [54]

    In: CVPR

    You, Q., Jin, H., Wang, Z., Fang, C., Luo, J.: Image captioning with semantic attention. In: CVPR. pp. 4651–4659 (2016)

    2016

  55. [55]

    arXiv preprint arXiv:2106.071663(7) (2021)

    Yu, Y., Wang, X., Hu, W., Luo, X., Li, C.: 2rd place solutions in the hc-stvg track of person in context challenge 2021. arXiv preprint arXiv:2106.071663(7) (2021)

    Pith/arXiv arXiv 2021

  56. [56]

    In: AAAI

    Zhang, S., Peng, H., Fu, J., Luo, J.: Learning 2d temporal adjacent networks for moment localization with natural language. In: AAAI. vol. 34, pp. 12870–12877 (2020)

    2020

  57. [57]

    In: CVPR

    Zhang, Y., Chen, X., Jia, J., Liu, S., Ding, K.: Text-visual prompting for efficient 2d temporal video grounding. In: CVPR. pp. 14794–14804 (2023)

    2023

  58. [58]

    arXiv preprint arXiv:2008.06941 (2020)

    Zhang, Z., Zhao, Z., Lin, Z., Huai, B., Yuan, N.J.: Object-aware multi- branch relation networks for spatio-temporal video grounding. arXiv preprint arXiv:2008.06941 (2020)

    Pith/arXiv arXiv 2008

  59. [59]

    In: CVPR

    Zhang, Z., Zhao, Z., Zhao, Y., Wang, Q., Liu, H., Gao, L.: Where does it exist: Spatio-temporal video grounding for multi-form sentences. In: CVPR. pp. 10668– 10677 (2020)

    2020

  60. [60]

    arXiv preprint arXiv:2602.13313 (2026)

    Zhao, H., Ong, Y.S., Zhou, J.T.: Agentic spatio-temporal grounding via collabo- rative reasoning. arXiv preprint arXiv:2602.13313 (2026)

    arXiv 2026

  61. [61]

    In: ECCV

    Zhu, C., Zhou, Y., Shen, Y., Luo, G., Pan, X., Lin, M., Chen, C., Cao, L., Sun, X., Ji, R.: Seqtr: A simple yet universal network for visual grounding. In: ECCV. pp. 598–615. Springer (2022)

    2022

  62. [62]

    arXiv preprint arXiv:2504.10479 (2025)

    Zhu, J., Wang, W., Chen, Z., Liu, Z., Ye, S., Gu, L., Tian, H., Duan, Y., Su, W., Shao, J., et al.: Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models. arXiv preprint arXiv:2504.10479 (2025)

    Pith/arXiv arXiv 2025

  63. [63]

    arXiv preprint arXiv:2010.04159 (2020)

    Zhu, X., Su, W., Lu, L., Li, B., Wang, X., Dai, J.: Deformable detr: Deformable transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159 (2020)

    Pith/arXiv arXiv 2010

This paper was first reviewed by deepseek-v4-flash on August 2, 2026.