Pith. sign in

REVIEW 3 major objections 5 minor 135 references

The paper claims that RGBT tracking improves when fusion and temporal-calibration parameters are generated per target from template features, using a target-adaptive hypernetwork, rather than fixed or router-selected.

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 ↗

PAFCNet dynamically generates target-conditioned parameters for multimodal fusion and spatio-temporal calibration in RGBT tracking, achieving competitive benchmark results.

T0 review reviewed 2026-08-04 challenge →

load-bearing objection Nice idea, but the headline numbers are selected from a test-set grid search, so the empirical claim needs a fix before it can be taken at face value. the 3 major comments →

arxiv 2608.01807 v1 pith:LMVTS64K submitted 2026-08-03 cs.CV

Parameter-Dynamic Adaptive Fusion and Calibration Network for RGBT Tracking

Zhaoding Ding , Chenglong Li , Jiandong Jin , Kewei Ying , Wentao Wu This is my paper
classification cs.CV
keywords RGBT trackingmultimodal fusionhypernetworkspatio-temporal calibrationvisual object trackingtarget-adaptive parametersthermal infrared tracking
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

RGBT tracking combines visible-light and thermal images of the same scene; existing trackers usually fuse them with a fixed function or by choosing among a few predefined fusion operations. The paper's claim is that fusion should be parameter-dynamic: a hypernetwork—a small network that produces parameters for another network—reads the target's template, aggregates its stable and recently changed appearance, and generates fusion and temporal-calibration weights for the current frame. This lets one tracker adapt to appearance changes and to one modality becoming unreliable, rather than applying the same fusion strategy to every target. The paper implements this in PAFCNet with a target-aware fusion module and a dynamic spatio-temporal calibration module, and reports leading results on GTOT, RGBT210, RGBT234, and LasHeR. If correct, generating per-target parameters is a workable alternative to both fixed fusion and router-based dynamic architectures.

Core claim

PAFCNet's core discovery is that the state of the target itself can be used as the conditioning signal for multimodal fusion in RGBT tracking. A Target-Adaptive Hypernetwork (TA-HyperNet) takes initial and dynamic template tokens, normalizes them, and aggregates them through a global content-aware branch and a multi-query local branch. It then outputs bounded channel-wise scaling and shifting parameters. One set of parameters modulates a Target-Aware Parameter-Dynamic Fusion Module (TPDFM) inserted in the 10th–12th transformer layers, where token-wise RGB and TIR weights are predicted with independent sigmoids so the two modalities do not compete. Another set of parameters calibrates histori

What carries the argument

The load-bearing object is TA-HyperNet, a small hypernetwork built around a Template Representation Aggregation Module (TRAM). TRAM condenses the template tokens into one global summary and K local summaries (K=4), maps them to a compact latent vector, and generates parameter pairs (γ1,β1) and (γ2,β2) via tanh, bounding the modulation to stabilize training. These parameters modulate the hidden features of TPDFM and DSCM in residual form x⊙(1+γ)+β, so the pretrained feature transformation is preserved while target-conditioned adjustment is added. TPDFM uses the modulated features to produce token-wise modality weights with independent sigmoids; DSCM uses a separate parameter-generation head t

Load-bearing premise

The reported LasHeR gains assume that the hyperparameters in Tables 4 and 5 (number of spatio-temporal tokens, TPDFM insertion layers) were chosen without using the LasHeR test set; those tables report test-set numbers for each choice.

What would settle it

Train PAFCNet and its strongest baseline from the same checkpoint, choosing token count and insertion layers on a held-out validation split of LasHeR before predicting on the test set. If the 1.8-point PR and 1.3-point SR advantages over UATrack shrink to run-to-run noise, the parameter-dynamic mechanism is not the source of the gain.

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

If this is right

  • The fusion module slots into the last three transformer layers of a one-stream ViT tracker, so existing pipelines can adopt it without redesigning the backbone.
  • Independent sigmoid weights let the fusion module amplify both RGB and TIR responses together or suppress both, which suits complementary and unreliable modalities.
  • Calibrating each modality's propagated history separately should reduce cross-modal noise contamination over long sequences; the paper's attribute results on occlusion and motion blur support this.
  • Ablations show each component carries part of the gain: template update, DSCM, and TPDFM each improve the metrics, so the paper claims the mechanism is additive.

Where Pith is reading between the lines

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

  • If the template-conditioning mechanism transfers, the same hypernetwork design could be applied to other multimodal tracking settings with a persistent identity reference, such as RGB-depth or audio-visual tracking; the paper only evaluates RGB-T.
  • The bounded tanh modulation and independent sigmoid weighting suggest robustness to missing or degraded modalities, but the paper does not run explicit modality-missing experiments; that is a direct test of the mechanism's advertised property.
  • A comparison against random or history-agnostic conditioning would isolate whether the gains come from using the template's current state rather than from simply adding parameters; the paper does not include that control.
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 / 5 minor

Summary. The paper proposes PAFCNet, an RGBT tracking architecture in which a Target-Adaptive Hypernetwork (TA-HyperNet) generates target-conditioned channel-wise modulation parameters from initial and dynamic template tokens. These parameters are used by a Target-aware Parameter-Dynamic Fusion Module (TPDFM) to modulate multimodal fusion, and by a Dynamic Spatio-temporal Calibration Module (DSCM) to calibrate propagated modality-specific spatio-temporal tokens. The model is built on an OSTrack-style baseline and evaluated on GTOT, RGBT210, RGBT234, and LasHeR, reporting state-of-the-art or competitive results on all four benchmarks. The central claim is that dynamically generating fusion and calibration parameters from template representations improves robustness to target appearance change and modality-quality fluctuation.

Significance. The idea of using a hypernetwork to generate per-target, per-frame modulation parameters for multimodal fusion and temporal calibration is timely and potentially useful for RGBT tracking. The architecture is clearly motivated, and the component ablations in Table 2 are monotonic and internally consistent. If the reported benchmark numbers are confirmed under unbiased model selection, PAFCNet would be a meaningful step beyond fixed-parameter fusion and dynamic-architecture routing. The paper also provides a useful comparison among static fusion, dynamic-architecture fusion, and parameter-dynamic fusion in Table 3. However, the empirical contribution is currently undermined by a model-selection protocol that uses the test sets directly for hyperparameter choice, and by missing reproducibility artifacts.

major comments (3)
  1. [§Experiments, Tables 4 and 5] The final configuration is selected on the benchmark test sets. The text states that the insertion-layer study was performed 'on RGBT234 and LasHeR test set' (Table 4) and the token-count study 'on the LasHeR test set' (Table 5), and the best settings (layers 10-12 and Np=64) are then used in Table 1. Therefore the reported LasHeR numbers are maxima over a model-selection grid, not unbiased estimates. The margins over the strongest complete competitor UATrack are small (PAFCNet-384: PR +1.8, NPR +1.7, SR +1.3; PAFCNet-256: PR +1.3, NPR +1.0, SR +0.8). Selection among four layer configurations and five token counts could plausibly account for these differences. The central performance claim requires evaluation on a held-out validation split before Table 1 is accepted.
  2. [§Ablation Studies / Implementation Details] No error bars or multiple-seed results are reported; all ablation and comparison numbers are single runs. Several improvements are within typical run-to-run variability (e.g., RGBT210 SR +0.1 over RAGTrack, GTOT MSR +0.6 over VCT), and the component-level gains in Table 2 are 1-2 percentage points on LasHeR. Given the small margins and the selection bias identified above, the paper should report variance over at least three seeds or otherwise justify statistical significance. This is secondary to the test-set selection issue but directly affects the empirical soundness of the ablations.
  3. [§Implementation Details and supplementary] The paper repeatedly defers essential details to 'the supplementary material' (training setup, 'other hyperparameters'), but no supplementary file is included with the arXiv submission. The dynamic template update rule is not specified, and the exact separation/re-initialization of propagated spatio-temporal tokens in DSCM is described only at a high level. Without code or a complete supplementary, the experiments cannot be independently reproduced. This is especially problematic because the method's main novelty depends on the precise conditioning of TA-HyperNet on template and token states.
minor comments (5)
  1. [Figure 3] The caption area contains garbled Unicode escape sequences (e.g., '/uni00000031/uni00000032'); the figure/caption needs cleanup before publication.
  2. [§Experiments, 'Effect of Insertion Layers'] The phrase 'on RGBT234 and LasHeR test set' should be reworded to 'test sets' or clarified. More importantly, the paper should state whether any validation split was used for RGBT234 or whether the full benchmark was used for selection.
  3. [Equation (15)] The notation 'TRAMP' is not defined; presumably it means TRAM with a modality-specific projection p. Please clarify.
  4. [Table 1] For GTOT, PAFCNet-384 has lower MSR (80.2) than PAFCNet-256 (81.1), while the text says the best MSR is 81.1. The narrative is correct but the table layout makes the comparison slightly confusing; consider highlighting the best per configuration.
  5. [References] Several references lack complete venue or page information (e.g., TUMFNet, RAGTrack, CADTrack), and a few entries are listed by arXiv identifier. Please normalize the bibliography.

Circularity Check

0 steps flagged

No derivation-level circularity; reported gains are empirical, with a test-set selection caveat that is a soundness risk.

full rationale

PAFCNet's central claim is an empirical performance comparison, not a derivation. The TA-HyperNet generates modulation parameters via the paper's own equations (Eqs. 5, 11, 16) from template tokens; these are architectural definitions, not identities that presuppose the benchmark results. The fusion weights in Eqs. 13-14 are predicted from the input features and target embedding, so the output is not a fitted constant in disguise. Ablations in Tables 2-5 compare variants trained with the same objective; they do not derive the final numbers from the hyperparameters themselves. The only concern near the circularity boundary is that Tables 4 and 5 select TPDFM insertion layers and the number of spatio-temporal tokens using the LasHeR test set, and Table 1 then reports the resulting LasHeR scores as evidence of competitiveness. That is a soundness/overfitting risk, not a derivation-equivalence circularity: no equation reduces to a fitted parameter, and no self-citation carries the argument. Self-citations are to related work and baselines, not to a uniqueness theorem or an adopted ansatz. Therefore, no significant circularity in the claimed derivation chain; the score reflects the test-set selection caveat.

Axiom & Free-Parameter Ledger

6 free parameters · 3 axioms · 3 invented entities

The central claim rests on the new modules (TA-HyperNet, TPDFM, DSCM) and on the chosen hyperparameters. Several hyperparameters (Np, insertion layers) were selected using test set performance, which is a fitting-like procedure that affects the reported gains.

free parameters (6)
  • Np (number of spatio-temporal tokens) = 64
    Chosen by ablation on LasHeR test set (Table 5); 64 gives best results.
  • TPDFM insertion layers = 10, 11, 12
    Chosen by ablation on LasHeR test set (Table 4); inserting into final three layers gives best results.
  • K (number of summary queries in TRAM) = 4
    Design choice reported in Methodology; not ablated.
  • H (hidden dimension for modulation parameters) = 256
    Design choice reported in Methodology; not ablated.
  • Cp (projected interaction dimension) = 64
    Design choice reported in Methodology; not ablated.
  • Dt (target embedding dimension) = 8
    Design choice reported in Methodology; not ablated.
axioms (3)
  • domain assumption OSTrack baseline and DropMAE pretrained weights provide a valid starting point
    The model is initialized with DropMAE pretrained weights and trained with the OSTrack objective; performance of the method depends on these external resources.
  • domain assumption Benchmark evaluation protocols (MPR, MSR, PR, SR, NPR) are correctly applied and comparable across methods
    The paper reports numbers from standard benchmarks (GTOT, RGBT210, RGBT234, LasHeR) and compares to published results without independent verification.
  • standard math Self-attention and softmax operations behave as standard
    The modules use standard transformer self-attention and softmax; no new mathematical claims are made.
invented entities (3)
  • TA-HyperNet (Target-Adaptive Hypernetwork) independent evidence
    purpose: Generates target-conditioned modulation parameters for fusion and calibration from template representations.
    Ablation studies show that including TA-HyperNet-based TPDFM and DSCM improves benchmark results, providing a falsifiable handle within the paper.
  • TPDFM (Target-aware Parameter-Dynamic Fusion Module) independent evidence
    purpose: Modulates multimodal fusion using hypernetwork-generated parameters, enabling target-state-adaptive fusion.
    Table 3 shows TPDFM outperforms static fusion and dynamic-architecture fusion on LasHeR.
  • DSCM (Dynamic Spatio-temporal Calibration Module) independent evidence
    purpose: Calibrates spatio-temporal tokens before propagation to suppress accumulated tracking noise.
    Table 2 shows adding DSCM improves results over template-update-only baseline.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Parameter-Dynamic Adaptive Fusion and Calibration Network for RGBT Tracking." pith.science (2026). https://pith.science/paper/LMVTS64K

@misc{pith2026260801807,
  author       = {Pith},
  title        = {Pith review of: Parameter-Dynamic Adaptive Fusion and Calibration Network for RGBT Tracking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LMVTS64K}},
  note         = {Machine review of arXiv:2608.01807}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Existing RGBT trackers typically employ fusion functions with fixed parameters across different targets and scenarios. Although dynamic-architecture methods improve fusion flexibility by selecting among predefined operations, they still cannot adapt the fusion parameters to the evolving target state. To address these issues, we propose a Parameter-Dynamic Adaptive Fusion and Calibration Network (PAFCNet) for RGBT tracking. PAFCNet dynamically generates target-conditioned parameters for multimodal fusion and temporal calibration, enabling the tracking process to adapt to target appearance variations and modality quality fluctuations. Specifically, we introduce a Target-Adaptive Hypernetwork (TA-HyperNet) that leverages template representations, which preserve stable target identity and recent appearance changes with less background interference, to generate target-conditioned parameters for subsequent fusion and calibration. Based on TA-HyperNet, we design a target-aware parameter-dynamic fusion module that uses the generated parameters to modulate the fusion process. This enables the fusion module to adapt to changes in target appearance and complex scene conditions. Furthermore, since spatio-temporal information propagation may accumulate tracking noise, we propose a dynamic spatio-temporal calibration module that employs TA-HyperNet to generate calibration parameters for spatio-temporal tokens. By dynamically calibrating historical information before propagation, the module improves the reliability of temporal representations. Experimental results demonstrate that PAFCNet achieves competitive performance on multiple RGBT tracking benchmarks.

Figures

Figures reproduced from arXiv: 2608.01807 by Chenglong Li, Jiandong Jin, Kewei Ying, Wentao Wu, Zhaoding Ding.

Figure 1
Figure 1. Figure 1: (a) Multimodal fusion with fixed parameters and a [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The overall framework of our method. TRAM denotes the template representation aggregation module. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Attribute-based evaluation on the LasHeR dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Attention evolution for RGB (top) and TIR (bot [PITH_FULL_IMAGE:figures/full_fig_p007_4.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

135 extracted references · 1 canonical work pages

  1. [1]

    Journal of Image and Graphics , volume=

    A benchmark dataset for high-altitude UAV multi-modal tracking , author=. Journal of Image and Graphics , volume=. 2025 , doi=

    2025

  2. [2]

    Proceedings of IEEE Conference on Computer Vision and Pattern Recognition , year=

    Single-Model and Any-Modality for Video Object Tracking , author=. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition , year=

  3. [3]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Sdstrack: Self-distillation symmetric adapter learning for multi-modal visual object tracking , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  4. [4]

    Proceedings of the AAAI Conference on Artificial Intelligence , year=

    Temporal Adaptive RGBT Tracking with Modality Prompt , author=. Proceedings of the AAAI Conference on Artificial Intelligence , year=

  5. [5]

    Proceedings of the 31st ACM International Conference on Multimedia , pages=

    Quality-Aware RGBT Tracking via Supervised Reliability Learning and Weighted Residual Guidance , author=. Proceedings of the 31st ACM International Conference on Multimedia , pages=

  6. [6]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Bridging Search Region Interaction With Template for RGB-T Tracking , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  7. [7]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Efficient RGB-T Tracking via Cross-Modality Distillation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  8. [8]

    IEEE Transactions on Intelligent Transportation Systems , pages=

    Dynamic fusion network for RGBT tracking , author=. IEEE Transactions on Intelligent Transportation Systems , pages=

  9. [9]

    IEEE Transactions on Multimedia , year=

    MFGNet: Dynamic modality-aware filter generation for RGB-T tracking , author=. IEEE Transactions on Multimedia , year=

  10. [10]

    Proceedings of the IEEE conference on computer vision and pattern recognition , year =

    Zhang Pengyu and Jie Zhao and Dong Wang and Huchuan Lu and Xiang Ruan , title =. Proceedings of the IEEE conference on computer vision and pattern recognition , year =

  11. [11]

    Proceedings of the 30th ACM International Conference on Multimedia , pages=

    Prompting for multi-modal tracking , author=. Proceedings of the 30th ACM International Conference on Multimedia , pages=

  12. [12]

    IEEE Transactions on Neural Networks and Learning Systems , year=

    Duality-gated mutual condition network for RGBT tracking , author=. IEEE Transactions on Neural Networks and Learning Systems , year=

  13. [13]

    IEEE Transactions on Image Processing , pages=

    RGBT tracking via multi-adapter network with hierarchical divergence loss , author=. IEEE Transactions on Image Processing , pages=

  14. [14]

    European conference on computer vision , pages=

    Challenge-aware RGBT tracking , author=. European conference on computer vision , pages=

  15. [15]

    Computational Visual Media , year=

    Wang, Wenhai and Xie, Enze and Li, Xiang and Fan, Deng-Ping and Song, Kaitao and Liang, Ding and Lu, Tong and Luo, Ping and Shao, Ling , title=. Computational Visual Media , year=. doi:10.1007/s41095-022-0274-8 , url=

    work page doi:10.1007/s41095-022-0274-8

  16. [16]

    MCMAE: Masked Convolution Meets Masked Autoencoders , url =

    Gao, Peng and Ma, Teli and Li, Hongsheng and Lin, Ziyi and Dai, Jifeng and Qiao, Yu , booktitle =. MCMAE: Masked Convolution Meets Masked Autoencoders , url =

  17. [17]

    2025 , isbn =

    Shi, Zican and Hu, Jing and Ren, Jie and Ye, Hengkang and Yuan, Xuyang and Ouyang, Yan and He, Jia and Ji, Bo and Guo, Junyu , title =. 2025 , isbn =. doi:10.1609/aaai.v39i7.32740 , booktitle =

    doi:10.1609/aaai.v39i7.32740 2025

  18. [18]

    SwinTrack: A Simple and Strong Baseline for Transformer Tracking , url =

    Lin, Liting and Fan, Heng and Zhang, Zhipeng and Xu, Yong and Ling, Haibin , booktitle =. SwinTrack: A Simple and Strong Baseline for Transformer Tracking , url =

  19. [19]

    Siamese Masked Autoencoders , url =

    Gupta, Agrim and Wu, Jiajun and Deng, Jia and Li, Fei-Fei , booktitle =. Siamese Masked Autoencoders , url =

  20. [20]

    IEEE Transactions on Intelligent Vehicles , pages=

    Zhu, Yabin and Li, Chenglong and Tang, Jin and Luo, Bin , title=. IEEE Transactions on Intelligent Vehicles , pages=

  21. [21]

    Proceedings of the 31st ACM International Conference on Multimedia , pages=

    Wang, Zeyu and Colonnier, Fabien and Zheng, Jinghong and Acharya, Jyotibdha and Jiang, Wenyu and Huang, Kejie , title=. Proceedings of the 31st ACM International Conference on Multimedia , pages=

  22. [22]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , pages=

    Wang, Xiaogang and Wang, Meng and Li, Wei , title=. IEEE Transactions on Pattern Analysis and Machine Intelligence , pages=

  23. [23]

    2017 IEEE International Conference on Robotics and Automation (ICRA) , pages=

    Chen, Long and Sun, Libo and Yang, Teng and Fan, Lei and Huang, Kai and Xuanyuan, Zhe , title=. 2017 IEEE International Conference on Robotics and Automation (ICRA) , pages=

    2017

  24. [24]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , pages=

    Gao, Chen and Liu, Si and Chen, Jinyu and Wang, Luting and Wu, Qi and Li, Bo and Tian, Qi , title=. IEEE Transactions on Pattern Analysis and Machine Intelligence , pages=

  25. [25]

    Chenglong Li and Tao Wang and Zhaodong Ding and Yun Xiao and Jin Tang , title=. , year=

  26. [26]

    Quality-Aware Spatio-Temporal Transformer Network for RGBT Tracking , year=

    Ding, Zhaodong and Li, Chenglong and Wang, Tao and Wang, Futian , journal=. Quality-Aware Spatio-Temporal Transformer Network for RGBT Tracking , year=

  27. [27]

    Information Fusion , pages =

    Zhangyong Tang and Tianyang Xu and Hui Li and Xiao-Jun Wu and XueFeng Zhu and Josef Kittler , title =. Information Fusion , pages =

  28. [28]

    IEEE Transactions on Intelligent Vehicles , pages=

    Yabin Zhu and Chenglong Li and Bin Luo and Jin Tang , title=. IEEE Transactions on Intelligent Vehicles , pages=

  29. [29]

    AAAI Conference on Artificial Intelligence , pages=

    Bing Cao and Junliang Guo and Pengfei Zhu and Qinghua Hu , title=. AAAI Conference on Artificial Intelligence , pages=

  30. [30]

    Proceedings of the AAAI Conference on Artificial Intelligence, 39(4)

    Xiantao Hu and Ying Tai and Xu Zhao and Chen Zhao and Zhenyu Zhang and Jun Li and Bineng Zhong and Jian Yang , title=. Proceedings of the AAAI Conference on Artificial Intelligence, 39(4). , year=

  31. [31]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Hui, Tianrui and Xun, Zizheng and Peng, Fengguang and Huang, Junshi and Wei, Xiaoming and Wei, Xiaolin and Dai, Jiao and Han, Jizhong and Liu, Si , title=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  32. [32]

    IEEE Transactions on Circuits and Systems for Video Technology , year=

    Zhu, Yabin and Li, Chenglong and Tang, Jin and Luo, Bin and Wang, Liang , title=. IEEE Transactions on Circuits and Systems for Video Technology , year=

  33. [33]

    IEEE Transactions on Circuits and Systems for Video Technology , year=

    Hu, Xiantao and Zhong, Bineng and Liang, Qihua and Zhang, Shengping and Li, Ning and Li, Xianxian , title=. IEEE Transactions on Circuits and Systems for Video Technology , year=

  34. [34]

    IEEE Transactions on Circuits and Systems for Video Technology , pages=

    Zhang, Tianlu and Liu, Xueru and Zhang, Qiang and Han, Jungong , title=. IEEE Transactions on Circuits and Systems for Video Technology , pages=

  35. [35]

    IEEE Transactions on Multimedia , pages=

    Wang, Xiao and Shu, Xiujun and Zhang, Shiliang and Jiang, Bo and Wang, Yaowei and Tian, Yonghong and Wu, Feng , title=. IEEE Transactions on Multimedia , pages=

  36. [36]

    IEEE Transactions on Neural Networks and Learning Systems , pages=

    Lu, Andong and Qian, Cun and Li, Chenglong and Tang, Jin and Wang, Liang , title=. IEEE Transactions on Neural Networks and Learning Systems , pages=

  37. [37]

    IEEE Transactions on Intelligent Transportation Systems , pages=

    Peng, Jingchao and Zhao, Haitao and Hu, Zhengwei , title=. IEEE Transactions on Intelligent Transportation Systems , pages=

  38. [38]

    2024 4th International Conference on Computer Science and Blockchain (CCSB) , pages=

    Ren, Chao and Liu, Ruian , title=. 2024 4th International Conference on Computer Science and Blockchain (CCSB) , pages=

    2024

  39. [39]

    2023 IEEE 6th International Conference on Electronics and Communication Engineering (ICECE) , pages=

    Shi, Siqi and Li, Nanting and Yin, Youwei , title=. 2023 IEEE 6th International Conference on Electronics and Communication Engineering (ICECE) , pages=

    2023

  40. [40]

    Proceedings of the AAAI Conference on Artificial Intelligence , pages =

    Shi, Liangtao and Zhong, Bineng and Liang, Qihua and Li, Ning and Zhang, Shengping and Li, Xianxian , title =. Proceedings of the AAAI Conference on Artificial Intelligence , pages =

  41. [41]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages =

    Xie, Jinxia and Zhong, Bineng and Mo, Zhiyi and Zhang, Shengping and Shi, Liangtao and Song, Shuxiang and Ji, Rongrong , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages =

  42. [42]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , pages =

    Yan, Bin and Peng, Houwen and Fu, Jianlong and Wang, Dong and Lu, Huchuan , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , pages =

  43. [43]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , pages=

    Li, Bo and Peng, Fengguang and Hui, Tianrui and Wei, Xiaoming and Wei, Xiaolin and Zhang, Lijun and Shi, Hang and Liu, Si , title=. IEEE Transactions on Pattern Analysis and Machine Intelligence , pages=

  44. [44]

    2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    Cui, Yutao and Jiang, Cheng and Wang, Limin and Wu, Gangshan , title=. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    2022

  45. [45]

    CoRR , year =

    Alexey Dosovitskiy and Lucas Beyer and Alexander Kolesnikov and Dirk Weissenborn and Xiaohua Zhai and Thomas Unterthiner and Mostafa Dehghani and Matthias Minderer and Georg Heigold and Sylvain Gelly and Jakob Uszkoreit and Neil Houlsby , title =. CoRR , year =

  46. [46]

    Hongyu Wang and Xiaotao Liu and Yifan Li and Meng Sun and Dian Yuan and Jing Liu , title=. Proceedings of the Thirty-Eighth AAAI Conference on Artificial Intelligence and Thirty-Sixth Conference on Innovative Applications of Artificial Intelligence and Fourteenth Symposium on Educational Advances in Artificial Intelligence , year=

  47. [47]

    ECCV , year=

    Ye, Botao and Chang, Hong and Ma, Bingpeng and Shan, Shiguang and Chen, Xilin , title=. ECCV , year=

  48. [48]

    Advances in Neural Information Processing Systems , pages =

    Lin, Liting and Fan, Heng and Zhang, Zhipeng and Xu, Yong and Ling, Haibin , title =. Advances in Neural Information Processing Systems , pages =

  49. [49]

    and Kaiser,

    Vaswani, Ashish and Shazeer, Noam and Parmar, Niki and Uszkoreit, Jakob and Jones, Llion and Gomez, Aidan N. and Kaiser,. Attention is all you need , booktitle =

  50. [50]

    2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    Wei, Xi and Zhang, Tianzhu and Li, Yan and Zhang, Yongdong and Wu, Feng , title=. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    2020

  51. [51]

    Wildes , title=

    Isma Hadji and Richard P. Wildes , title=. arXiv preprint arXiv:1803.08834 , year=

    Pith/arXiv arXiv

  52. [52]

    arXiv e-prints , pages =

    Andong Lu and Wanyu Wang and Chenglong Li and Jin Tang and Bin Luo , title=. arXiv e-prints , pages =

  53. [53]

    IEEE Transactions on Circuits and Systems for Video Technology , year=

    Sun, Dengdi and Pan, Yajie and Lu, Andong and Li, Chenglong and Luo, Bin , title=. IEEE Transactions on Circuits and Systems for Video Technology , year=

  54. [54]

    2012 IEEE Conference on Computer Vision and Pattern Recognition , pages=

    Jia, Xu and Lu, Huchuan and Yang, Ming-Hsuan , title=. 2012 IEEE Conference on Computer Vision and Pattern Recognition , pages=

    2012

  55. [55]

    Template Shift and Background Suppression for Visual Object Tracking

    Wu, Yiwei and Qi, Ke and Chen, Wenbin and Zhang, Jingdong and Qi, Yutao. Template Shift and Background Suppression for Visual Object Tracking. Knowledge Science, Engineering and Management. 2023

    2023

  56. [56]

    IEEE Access , pages=

    Tian, Erlin and Lei, Yunpeng and Sun, Junfeng and Zhou, Keyan and Zhou, Bin and Li, Hanfei , title=. IEEE Access , pages=

  57. [57]

    Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence,

    Zhihong Fu and Zehua Fu and Qingjie Liu and Wenrui Cai and Yunhong Wang , title=. Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence,

  58. [58]

    arXiv preprint arXiv:2201.08949 , year=

    Tang, Zhangyong and Xu, Tianyang and Wu, Xiao-Jun , title=. arXiv preprint arXiv:2201.08949 , year=

    Pith/arXiv arXiv

  59. [59]

    Proceedings of the AAAI Conference on Artificial Intelligence, 38(5) , pages =

    Liangtao Shi and Bineng Zhong and Qihua Liang and Ning Li and Shengping Zhang and Xianxian Li , title=. Proceedings of the AAAI Conference on Artificial Intelligence, 38(5) , pages =

  60. [60]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages =

    Jinxia Xie and Bineng Zhong and Zhiyi Mo and Shengping Zhang and Liangtao Shi and Shuxiang Song and Rongrong Ji , title=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages =

  61. [61]

    IEEE Transactions on Circuits and Systems for Video Technology , year=

    Xue, Wanli and Xu, Chao and Feng, Zhiyong , title=. IEEE Transactions on Circuits and Systems for Video Technology , year=

  62. [62]

    IEEE Transactions on Multimedia , pages=

    Mai, Sijie and Zeng, Ying and Hu, Haifeng , title=. IEEE Transactions on Multimedia , pages=

  63. [63]

    2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Hu, Jie and Shen, Li and Sun, Gang , title=. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    2018

  64. [64]

    2019 IEEE/CVF International Conference on Computer Vision (ICCV) , pages=

    Xizhou Zhu and Dazhi Cheng and Zheng Zhang and Stephen Lin and Jifeng Dai , title=. 2019 IEEE/CVF International Conference on Computer Vision (ICCV) , pages=

    2019

  65. [65]

    2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW) , pages=

    Lichao Zhang and Martin Danelljan and Abel Gonzalez-Garcia and Joost van de Weijer and Fahad Shahbaz Khan , title=. 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW) , pages=

    2019

  66. [66]

    Signal Processing: Image Communication , pages =

    Xingchen Zhang and Ping Ye and Shengyun Peng and Jun Liu and Gang Xiao , title =. Signal Processing: Image Communication , pages =

  67. [67]

    IEEE Transactions on Circuits and Systems for Video Technology , pages=

    Li, Chenglong and Zhu, Chengli and Zhang, Jian and Luo, Bin and Wu, Xiaohao and Tang, Jin , title=. IEEE Transactions on Circuits and Systems for Video Technology , pages=

  68. [68]

    The IEEE International Conference on Computer Vision (ICCV) , pages=

    Zhang, Lichao and Gonzalez-Garcia, Abel and Weijer, Joost van de and Danelljan, Martin and Khan, Fahad Shahbaz , title =. The IEEE International Conference on Computer Vision (ICCV) , pages=

  69. [69]

    IEEE Transactions on Image Processing , pages=

    Li, Chenglong and Xue, Wanlin and Jia, Yaqing and Qu, Zhichen and Luo, Bin and Tang, Jin and Sun, Dengdi , title=. IEEE Transactions on Image Processing , pages=

  70. [70]

    Chan , title=

    Qiangqiang Wu and Tianyu Yang and Ziquan Liu and Baoyuan Wu and Ying Shan and Antoni B. Chan , title=. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    2023

  71. [71]

    International Conference on Learning Representations , year=

    Ilya Loshchilov and Frank Hutter , title=. International Conference on Learning Representations , year=

  72. [72]

    Pattern Recognition , pages =

    Chenglong Li and Xinyan Liang and Yijuan Lu and Nan Zhao and Jin Tang , title=. Pattern Recognition , pages =

  73. [73]

    Proceedings of the 25th ACM International Conference on Multimedia , pages =

    Li, Chenglong and Zhao, Nan and Lu, Yijuan and Zhu, Chengli and Tang, Jin , title =. Proceedings of the 25th ACM International Conference on Multimedia , pages =

  74. [74]

    Symmetry , pages=

    Zhang, Qian and Wang, Zihao and Liang, Hong , title=. Symmetry , pages=

  75. [75]

    2024 36th Chinese Control and Decision Conference (CCDC) , pages=

    Zhang, Tianshuo and Zhu, Linlin and Lin, Feng , title=. 2024 36th Chinese Control and Decision Conference (CCDC) , pages=

    2024

  76. [76]

    IEEE Transactions on Image Processing , pages=

    Li, Chenglong and Cheng, Hui and Hu, Shiyi and Liu, Xiaobai and Tang, Jin and Lin, Liang , title=. IEEE Transactions on Image Processing , pages=

  77. [77]

    AAAI Conference on Artificial Intelligence , pages=

    Yun Xiao and Mengmeng Yang and Chenglong Li and Lei Liu and Jin Tang , title=. AAAI Conference on Artificial Intelligence , pages=

  78. [78]

    IEEE Transactions on Neural Networks and Learning Systems , year=

    Lu, Andong and Qian, Cun and Li, Chenglong and Tang, Jin and Wang, Liang , title=. IEEE Transactions on Neural Networks and Learning Systems , year=

  79. [79]

    Proceedings of the 30th ACM International Conference on Multimedia , pages =

    Jinyu Yang and Zhe Li and Feng Zheng and Aleš Leonardis and Jingkuan Song , title=. Proceedings of the 30th ACM International Conference on Multimedia , pages =

  80. [80]

    2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    Pengyu Zhang and Jie Zhao and Dong Wang and Huchuan Lu and Xiang Ruan , title=. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages=

    2022

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.