Pith. sign in

REVIEW 3 major objections 6 minor 68 references

Dynamic Pattern Alignment Learning for Pretraining Lightweight Human-Centric Vision Models

T0 review · 3 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Distilling three visual patterns from a large teacher into a 5M-parameter student yields human-centric models that rival teachers tens to hundreds of times larger, using 1M unlabeled images the teacher never saw.

desk verdict A genuinely new pattern-decomposition distillation method with solid experiments, but the multi-person relation term is overclaimed and its evidential support is thin. read the letter →

arxiv 2508.07144 v1 pith:PYK6IZ5C submitted 2025-08-10 cs.CV

Xuanhan Wang , Huimin Deng , Ke Liu , Jun Wang , Lianli Gao , Jingkuan Song This is my paper
classification cs.CV
keywords knowledgedistillationhuman-centricvisionlightweighttransformersvisualpatternalignmentmixtureofexpertsself-supervisedpretrainingmulti-personinteractionrepresentationlearning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to show that a lightweight human-centric vision model can inherit the generalization of a large pretrained model without access to that model's private training data. Its method, DPAL, distills three kinds of visual knowledge—global identity, local body shape, and multi-person interaction—through a dynamic mixture-of-experts decoder and three alignment losses at image, pixel, and relation levels. Pretraining a 5M-parameter ViT-Tiny on about one million unlabeled images from a public collection for one day on eight GPUs is reported to match or approach the downstream accuracy of teacher and baseline models with 84M or 307M parameters across single-person, dense-prediction, multi-person, and cross-domain benchmarks. If correct, this gives a practical route to small, deployable human-centric models without large compute or restricted datasets.

What carries the argument

The load-bearing mechanism is the Dynamic Pattern Decoder (D-PaDe), a mixture-of-experts module whose three expert FFNs are not fixed but generated on the fly by a dynamic expert generator. Three learnable pattern queries pass through self- and cross-attention over the student's visual tokens to produce expert parameters, and a router assigns per-token expert weights. This decoder extracts exactly the three patterns the alignment losses use: a global identity pattern from multi-view single-person crops ($\ell_g$, MSE on global tokens), a local shape pattern from patch tokens masked by a teacher-derived body mask ($\ell_l$, MSE on local tokens), and a multi-person interaction pattern as a pat

What would settle it

Pretrain DPAL identically but replace the copy-paste multi-person images with real crowd images (for example, from CrowdHuman) in the relation-level loss, then compare on multi-person benchmarks such as CIHP and CrowdHuman. If the real-image variant performs no better than the synthetic variant, the copy-paste transfer assumption is not doing the work; if removing $\ell_r$ altogether leaves multi-person performance unchanged, the relation-level alignment is not load-bearing.

Watch

Extended reading notes

Core claim

At its core, DPAL claims that the generalization of a large human-centric vision model can be transferred to a lightweight student by decomposing the teacher's outputs into three patterns and forcing the student to match each one with its own expert pathway. The teacher (PATH-B) contributes global image tokens, local patch tokens, and attention maps. The student's Dynamic Pattern Decoder generates three expert networks on the fly, conditioned on the input image and three pattern queries, and a router activates the relevant expert per token. Three losses then align the student and teacher: a global-level MSE on identity tokens, a local-level MSE on body-masked patch tokens, and a relation-lev

Load-bearing premise

The multi-person alignment assumes synthetic copy-paste images transfer enough real-world interaction structure (occlusion, contact, spatial overlap) to teach patterns that generalize to real crowded scenes.

Editorial extensions

If this is right

  • Deployable backbones: at inference only the 5M student encoder remains, so human-centric tasks such as ReID, pose, parsing, and detection can run on resource-constrained devices with large-model-level accuracy.
  • Data-free upgrading: the student is trained on images the teacher has not seen, so a new or improved teacher can be swapped in without re-collecting or re-licensing the teacher's pretraining data.
  • Cheap retraining: pretraining from scratch on roughly 1M unlabeled images for 100 epochs on 8 GPUs makes the method reproducible by labs that cannot afford large 1000-GPU training runs.
  • A concrete distillation template: the three-pattern decomposition gives a target structure—global tokens, body-masked patches, and patch-correlation matrices—that other distillation schemes could adopt for human-centric models.

Reading between the lines

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

  • The paper does not test this, but the relation-level alignment uses only copy-paste composites; adding real multi-person images with occlusion and contact is a natural next test, and if it improves interaction-heavy benchmarks, the current results would underestimate what relation alignment can capture.
  • The same three-pattern recipe could be applied to non-human object categories (e.g., animals or vehicles) by replacing the body-shape mask with a class-agnostic foreground mask, making the approach a general pattern-distillation template rather than a human-specific one.
  • The saturation observed beyond 1M images was measured with a fixed teacher and distribution; a stronger teacher or harder data distribution could shift the scaling curve, so '1M is enough' should not be read as a universal law.
  • Since D-PaDe is discarded after pretraining, an ablation replacing it with a static three-branch decoder would isolate whether the dynamic expert generation itself, rather than just the three losses, carries the benefit.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes DPAL, a distillation-based pretraining framework for transferring human-centric visual knowledge from large vision models to lightweight students. The method identifies three visual patterns—global identity, local shape, and multi-person interaction—and introduces a dynamic pattern decoder (D-PaDe) with three dynamically generated experts to extract them. Three alignment losses are used: global-level MSE, local-level MSE, and relation-level KL divergence. Pretraining is performed on roughly one million unlabeled images from LUPerson that are not seen by the teacher, with multi-person images synthesized by copy-paste. The student is then fine-tuned on 15 (sometimes stated as 12) downstream datasets covering ReID, attribute recognition, pose/parsing, detection, and cross-domain tasks. The authors report substantial gains over prior self-supervised and distillation-based pretraining methods, and claim that a 5M ViT-Ti student approaches the generalization of much larger models such as PATH-B and Sapiens-L.

Significance. If the empirical claims hold, DPAL would be a practically valuable contribution: it offers a low-cost way to obtain generalizable lightweight human-centric models without accessing private pretraining datasets, and it directly addresses inter-pattern conflict via a dynamic expert architecture. The paper is generally well structured, and the ablations in Tables 3, 4, and 9 provide useful evidence for the individual losses and decoder design. The central claim is concrete and falsifiable. However, the relation-level pattern learning, which is one of the three advertised contributions, is supported by synthetic copy-paste data and mixed ablations; the comparison protocol with prior methods is not fully specified; and several core components are underspecified. These issues need to be resolved before the central claims can be accepted.

major comments (3)
  1. [§4.1, Eq. (6), Table 3] The relation-level alignment is trained exclusively on copy-paste synthetic multi-person images. The teacher's attention matrix on such composites cannot be assumed to encode genuine occlusion, contact, or crowd interaction. The evidence in Table 3 is mixed: ℓr alone yields the lowest detection AP (86.4), and adding ℓr to ℓl slightly decreases human parsing (55.7→55.5) and detection (88.4→88.1). The claimed transfer to real multi-person tasks (CrowdHuman/CIHP) therefore needs a direct test: use real multi-person images or a controlled synthetic-vs-real pretraining comparison, plus diagnostics on real crowded images. As written, the third advertised pattern is not convincingly established.
  2. [§4.3, Tables 1–2, §B.3] Comparisons in Tables 1–2 are not demonstrably controlled. The paper does not state which baseline numbers were reproduced with the same fine-tuning codebases/protocols listed in Table 8, nor whether multiple seeds/error bars were used. The abstract's 'large margin' claim over Proteus/TinyMIM and the 'similar to PATH-B/Sapiens-L' claim rest on these numbers. Please clarify provenance of every baseline result, re-run at least the strongest competitors under the same protocol, and report variance.
  3. [§3.1.2, Eqs. (2) and (4), §B.1] Two components central to the pattern-specific claims are underspecified. The local-shape mask M_shape in Eq. (2) has no extraction or threshold procedure and no ablation; the global-level M multi-view images in Eq. (4) have no construction description and M is not ablated. These free parameters affect the reported gains (Tables 3–4). Please provide exact definitions and ablations so the contributions are attributable and the method reproducible.
minor comments (6)
  1. [Abstract / §4.1 / §5 / Fig. 1 / Table 8] Inconsistent dataset counts: the abstract and §4.1 say 15 datasets, while Fig. 1 caption, §5, and Table 8 say 12. Please reconcile the count and make the table captions consistent.
  2. [Table 2(f)] Typo: Proteus AP-10K result '68/1' should be '68.1'. Also 'PETAzs' should be 'PETA' in several places.
  3. [§4.1 and Table 5] Table 5 lists '# imgs 1.2M' for DPAL, while §4.1 says LUP1M is a 1M subset. Clarify whether synthesized multi-person images are counted separately and why the number is 1.2M.
  4. [Figure 4] The x-axis is labeled 'Ratio of pretraining dataset' but the actual subset sizes (0.2M/0.5M/1M/2M/4M) are not marked, making the saturation claim difficult to verify. Please add tick labels or a clear scale.
  5. [Eqs. (1)–(3)] Notation is inconsistent: fF, fF_ls, gF_ls2, fF_rs mix math fonts and subscripts. Please standardize variable names for the student/teacher features.
  6. [References] Refs [14] and [15] are duplicates; Refs [10] and [11] are duplicates; Refs [35] and [36] are also duplicates. Please correct.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found; DPAL's claims rest on external downstream evaluation, not on constructed equivalences.

full rationale

DPAL is a knowledge-distillation pipeline: a frozen teacher (PATH-B) provides global, local, and relation-level features, and the student is trained to match them via Eqs. 4-6. This is the intended mechanism of distillation, not a hidden circularity: the student's features are regressed onto the teacher's features, and the resulting representations are then evaluated on external downstream benchmarks (Market1501, MSMT17, COCO, CrowdHuman, CIHP, etc.). No downstream metric is fed back into the pretraining objective, so the reported generalizability is not forced by construction. The paper's self-citations (SAIP [54], CPI-Parser [53], KE-RCNN [55]) are used as comparison baselines or downstream evaluation codebases; they are not load-bearing evidence for the central pattern-alignment claim. The use of copy-paste synthetic multi-person images (Sec. 4.1) may limit how well real multi-person interaction patterns are transferred, but this is a data-validity limitation, not a circular derivation: the relation-level loss still transfers whatever interaction signal the teacher provides on those images, and its contribution is separately ablated in Table 3. No uniqueness theorem, ansatz-smuggling citation, or renaming of a known result as a new derivation was found. The central empirical claim is therefore self-contained against external benchmarks, and the appropriate circularity score is 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 2 invented entities

The method rests on the taxonomical assumption that three patterns exhaust human-centric perception, on the reliability of the teacher's attention mask as a shape prior, and on the sufficiency of synthetic multi-person images. The invented module D-PaDe and pattern queries are validated only by internal ablations.

free parameters (4)
  • M_shape extraction (threshold/binarization) = unspecified
    The coarse foreground mask in Eq. 2 is derived from the teacher, but the exact extraction procedure and any thresholding are not specified, leaving a hand-tuned choice.
  • Number of multi-view images M (Eq. 4) = unspecified
    The global alignment loss averages over M views, but M is not reported.
  • KL divergence temperature (Eq. 6) = not reported (default 1 assumed)
    The relation-level alignment uses KL divergence without a stated temperature.
  • Loss weights = all 1.0
    The total loss is a sum of three equal-weight terms; no weighting search is reported.
assumptions (4)
  • domain assumption Human-centric perception decomposes into exactly three patterns: global identity, local shape, multi-person interaction.
    Introduction and Section 3.1.2; if other patterns (e.g., texture, temporal dynamics) are needed, the method would miss them.
  • domain assumption The teacher's attention-derived mask M_shape reliably localizes body shape.
    Eq. 2 uses M_shape to gate local tokens; the mask's reliability on unseen images is assumed.
  • domain assumption Copy-paste synthetic multi-person images capture sufficient interaction structure for relation alignment.
    Section 4.1: multi-person images are synthesized by copy-paste from single-person samples; real-world interactions (occlusion, contact) are not guaranteed.
  • standard math Standard transformer and MoE components (self-attention, cross-attention, FFN) are effective function approximators for alignment.
    Implicit in using neural networks for distillation; standard background.
invented entities (2)
  • Dynamic Pattern Decoder (D-PaDe)
    purpose: Adaptively routes student features to pattern-specific experts to avoid inter-pattern conflict.
    Only internal ablations (Tables 4 and 9) support its benefit; no external validation beyond this paper.
  • Pattern queries T_e
    purpose: Condition the expert generation on the desired pattern (global/local/relation).
    Part of the proposed architecture; no independent test outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dynamic Pattern Alignment Learning for Pretraining Lightweight Human-Centric Vision Models." pith.science (2026). https://pith.science/paper/PYK6IZ5C

@misc{pith2026250807144,
  author       = {Pith},
  title        = {Pith review of: Dynamic Pattern Alignment Learning for Pretraining Lightweight Human-Centric Vision Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PYK6IZ5C}},
  note         = {Machine review of arXiv:2508.07144}
}
read the original abstract

Human-centric vision models (HVMs) have achieved remarkable generalization due to large-scale pretraining on massive person images. However, their dependence on large neural architectures and the restricted accessibility of pretraining data significantly limits their practicality in real-world applications. To address this limitation, we propose Dynamic Pattern Alignment Learning (DPAL), a novel distillation-based pretraining framework that efficiently trains lightweight HVMs to acquire strong generalization from large HVMs. In particular, human-centric visual perception are highly dependent on three typical visual patterns, including global identity pattern, local shape pattern and multi-person interaction pattern. To achieve generalizable lightweight HVMs, we firstly design a dynamic pattern decoder (D-PaDe), acting as a dynamic Mixture of Expert (MoE) model. It incorporates three specialized experts dedicated to adaptively extract typical visual patterns, conditioned on both input image and pattern queries. And then, we present three levels of alignment objectives, which aims to minimize generalization gap between lightweight HVMs and large HVMs at global image level, local pixel level, and instance relation level. With these two deliberate designs, the DPAL effectively guides lightweight model to learn all typical human visual patterns from large HVMs, which can generalize to various human-centric vision tasks. Extensive experiments conducted on 15 challenging datasets demonstrate the effectiveness of the DPAL. Remarkably, when employing PATH-B as the teacher, DPAL-ViT/Ti (5M parameters) achieves surprising generalizability similar to existing large HVMs such as PATH-B (84M) and Sapiens-L (307M), and outperforms previous distillation-based pretraining methods including Proteus-ViT/Ti (5M) and TinyMiM-ViT/Ti (5M) by a large margin.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

68 extracted references · 56 canonical work pages

  1. [1]

    Mathilde Caron, Hugo Touvron, Ishan Misra, Hervé Jégou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. 2021. Emerging properties in self-supervised vision transformers. In Proceedings of the IEEE/CVF international conference on computer vision. 9650–9660. 13 Table 10: Impact of model architecture. We employ PATH-B as teacher model and perform distill...

    work page 2021

  2. [2]

    Hanting Chen, Yunhe Wang, Chang Xu, Zhaohui Yang, Chuanjian Liu, Boxin Shi, Chunjing Xu, Chao Xu, and Qi Tian. 2019. Data-free learning of student networks. In Proceedings of the IEEE/CVF international conference on computer vision. 3514–3522

    work page 2019

  3. [3]

    Weihua Chen, Xianzhe Xu, Jian Jia, Hao Luo, Yaohua Wang, Fan Wang, Rong Jin, and Xiuyu Sun. 2023. Beyond appearance: a semantic controllable self-supervised learning framework for human-centric visual tasks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 15050–15061

    work page 2023

  4. [4]

    Xianing Chen, Qiong Cao, Yujie Zhong, Jing Zhang, Shenghua Gao, and Dacheng Tao. 2022. Dearkd: data-efficient early knowledge distillation for vision transformers. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 12052–12062

    work page 2022

  5. [5]

    Hongsuk Choi, Gyeongsik Moon, JoonKyu Park, and Kyoung Mu Lee. 2022. Learning to estimate robust 3d human mesh from in-the-wild crowded scenes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 1475–1484. 14 Table 11: Impact of teacher size. We use ViT-Tiny as the student model and perform DPAL distillation separately w...

    work page 2022

  6. [6]

    Xuangeng Chu, Anlin Zheng, Xiangyu Zhang, and Jian Sun. 2020. Detection in Crowded Scenes: One Proposal, Multiple Predictions. InProceedings of the IEEE conference on computer vision and pattern recognition

    work page 2020

  7. [7]

    Yuanzheng Ci, Yizhou Wang, Meilin Chen, Shixiang Tang, Lei Bai, Feng Zhu, Rui Zhao, Fengwei Yu, Donglian Qi, and Wanli Ouyang. 2023. Unihcp: A unified model for human- centric perceptions. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 17840–17852

    work page 2023

  8. [8]

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. 2009. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition. Ieee, 248–255

    work page 2009

Show all 68 references
  1. [9]

    Yubin Deng, Ping Luo, Chen Change Loy, and Xiaoou Tang. 2014. Pedestrian attribute recognition at far distance. In Proceedings of the 22nd ACM international conference on Multimedia. 789–792

    2014

  2. [11]

    Zefeng Ding, Changxing Ding, Zhiyin Shao, and Dacheng Tao. 2021. Semantically self-aligned network for text-to-image part-aware person re-identification. arXiv preprint arXiv:2107.12666 (2021)

    2021 arXiv

  3. [12]

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al

  4. [13]

    Jiawei Fan, Chao Li, Xiaolong Liu, and Anabang Yao. 2024. ScaleKD: Strong Vision Transform- ers Could Be Excellent Teachers. Thirty-eighth Conference on Neural Information Processing Systems (2024)

    2024

  5. [14]

    Dengpan Fu, Dongdong Chen, Jianmin Bao, Hao Yang, Lu Yuan, Lei Zhang, Houqiang Li, and Dong Chen. 2021. Unsupervised pre-training for person re-identification. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 14750–14759

    2021

  6. [15]

    Dengpan Fu, Dongdong Chen, Jianmin Bao, Hao Yang, Lu Yuan, Lei Zhang, Houqiang Li, and Dong Chen. 2021. Unsupervised pre-training for person re-identification. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 14750–14759. 15 (b) Local shape...

    2021

  7. [16]

    Dengpan Fu, Dongdong Chen, Hao Yang, Jianmin Bao, Lu Yuan, Lei Zhang, Houqiang Li, Fang Wen, and Dong Chen. 2022. Large-scale pre-training for person re-identification with noisy labels. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2476–2486

    2022

  8. [17]

    Golnaz Ghiasi, Yin Cui, Aravind Srinivas, Rui Qian, Tsung-Yi Lin, Ekin D Cubuk, Quoc V Le, and Barret Zoph. 2021. Simple copy-paste is a strong data augmentation method for instance segmentation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognit...

    2021

  9. [18]

    Ke Gong, Xiaodan Liang, Yicheng Li, Yimin Chen, Ming Yang, and Liang Lin. 2018. Instance- level human parsing via part grouping network. In European Conference on Computer Vision. 770–785

    2018

  10. [19]

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, and Ross Girshick. 2022. Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 16000–16009

    2022

  11. [20]

    Shuting He, Hao Luo, Pichao Wang, Fan Wang, Hao Li, and Wei Jiang. 2021. Transreid: Transformer-based object re-identification. In Proceedings of the IEEE/CVF international conference on computer vision. 15013–15022

    2021

  12. [21]

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. 2015. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531 (2015)

    2015 arXiv

  13. [22]

    Wei Huang, Zhiliang Peng, Li Dong, Furu Wei, Jianbin Jiao, and Qixiang Ye. 2023. Generic-to- specific distillation of masked autoencoders. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 15996–16005

    2023

  14. [23]

    Jian Jia, Naiyu Gao, Fei He, Xiaotang Chen, and Kaiqi Huang. 2022. Learning disentangled attribute representations for robust pedestrian attribute recognition. In Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 36. 1069–1077

    2022

  15. [24]

    Jian Jia, Houjing Huang, Xiaotang Chen, and Kaiqi Huang. 2021. Rethinking of pedestrian attribute recognition: A reliable evaluation under zero-shot pedestrian identity setting. arXiv preprint arXiv:2107.03576 (2021)

    2021 arXiv

  16. [25]

    Menglin Jia, Mengyun Shi, Mikhail Sirotenko, Yin Cui, Claire Cardie, Bharath Hariharan, Hartwig Adam, and Serge Belongie. 2020. Fashionpedia: Ontology, segmentation, and an attribute localization dataset. In European Conference on Computer Vision. 316–332

    2020

  17. [26]

    Ding Jiang and Mang Ye. 2023. Cross-Modal Implicit Relation Reasoning and Aligning for Text-to-Image Person Retrieval. In Proceedings of the IEEE conference on computer vision and pattern recognition

    2023

  18. [27]

    Sheng Jin, Lumin Xu, Jin Xu, Can Wang, Wentao Liu, Chen Qian, Wanli Ouyang, and Ping Luo

  19. [28]

    Xuan Ju, Ailing Zeng, Jianan Wang, Qiang Xu, and Lei Zhang. 2023. Human-Art: A Versatile Human-Centric Dataset Bridging Natural and Artificial Scenes. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    2023

  20. [29]

    In European Conference on Computer Vision

    Whole-body human pose estimation in the wild. In European Conference on Computer Vision. Springer, 196–214

  21. [30]

    Peike Li, Yunqiu Xu, Yunchao Wei, and Yi Yang. 2020. Self-correction for human parsing. IEEE Transactions on Pattern Analysis and Machine Intelligence44, 6 (2020), 3260–3271

    2020

  22. [31]

    Rawal Khirodkar, Timur Bagautdinov, Julieta Martinez, Su Zhaoen, Austin James, Peter Selednik, Stuart Anderson, and Shunsuke Saito. 2024. Sapiens: Foundation for human vision models. In European Conference on Computer Vision. Springer, 206–228

    2024

  23. [32]

    Xiaodan Liang, Ke Gong, Xiaohui Shen, and Liang Lin. 2018. Look into person: Joint body parsing & pose estimation network and a new benchmark.IEEE transactions on pattern analysis and machine intelligence 41, 4 (2018), 871–885

    2018

  24. [33]

    Shuang Li, Tong Xiao, Hongsheng Li, Bolei Zhou, Dayu Yue, and Xiaogang Wang. 2017. Person search with natural language description. arXiv preprint arXiv:1702.05729

    2017 arXiv

  25. [34]

    Xihui Liu, Haiyu Zhao, Maoqing Tian, Lu Sheng, Jing Shao, Shuai Yi, Junjie Yan, and Xiaogang Wang. 2017. Hydraplus-net: Attentive deep features for pedestrian analysis. In Proceedings of the IEEE international conference on computer vision. 350–359. 17

    2017

  26. [35]

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C Lawrence Zitnick. 2014. Microsoft coco: Common objects in context. In European Conference on Computer Vision. Springer, 740–755

    2014

  27. [36]

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. 2021. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision. 10012–10022

    2021

  28. [38]

    Ilya Loshchilov and Frank Hutter. 2017. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017)

    2017 arXiv

  29. [39]

    Ilya Loshchilov and Frank Hutter. 2016. Sgdr: Stochastic gradient descent with warm restarts. arXiv preprint arXiv:1608.03983 (2016)

    2016 arXiv

  30. [40]

    Xiaoxuan Ma, Stephan Kaufhold, Jiajun Su, Wentao Zhu, Jack Terwilliger, Andres Meza, Yixin Zhu, Federico Rossano, and Yizhou Wang. 2023. Chimpact: A longitudinal dataset for understanding chimpanzee behaviors. Advances in Neural Information Processing Systems 36 (2023), 27501–27531

    2023

  31. [41]

    Hao Luo, Pichao Wang, Yi Xu, Feng Ding, Yanxin Zhou, Fan Wang, Hao Li, and Rong Jin

  32. [42]

    Yang Qian, Yinan Sun, Ali Kargarandehkordi, Parnian Azizian, Onur Cezmi Mutlu, Saimourya Surabhi, Pingyi Chen, Zain Jabbar, Dennis Paul Wall, and Peter Washington. 2024. Advancing Human Action Recognition with Foundation Models trained on Unlabeled Public Videos. arXiv preprin...

    2024 arXiv

  33. [43]

    Sucheng Ren, Fangyun Wei, Zheng Zhang, and Han Hu. 2023. Tinymim: An empirical study of distilling mim pre-trained models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 3687–3697

    2023

  34. [44]

    Namuk Park, Wonjae Kim, Byeongho Heo, Taekyung Kim, and Sangdoo Yun. 2023. What do self-supervised vision transformers learn? arXiv preprint arXiv:2305.00729 (2023)

    2023 arXiv

  35. [45]

    Shuai Shao, Zijian Zhao, Boxun Li, Tete Xiao, Gang Yu, Xiangyu Zhang, and Jian Sun. 2018. Crowdhuman: A benchmark for detecting human in a crowd. arXiv preprint arXiv:1805.00123 (2018)

    2018 arXiv

  36. [46]

    Zhiyin Shao, Xinyu Zhang, Changxing Ding, Jian Wang, and Jingdong Wang. 2023. Unified pre-training with pseudo texts for text-to-image person re-identification. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 11174–11184

    2023

  37. [47]

    Jinghuan Shang, Karl Schmeckpeper, Brandon B May, Maria Vittoria Minniti, Tarik Kelestemur, David Watkins, and Laura Herlant. 2024. Theia: Distilling diverse vision foundation models for robot learning. arXiv preprint arXiv:2407.20179 (2024)

    2024 arXiv

  38. [48]

    Seungwoo Son, Jegwang Ryu, Namhoon Lee, and Jaeho Lee. 2024. The Role of Masking for Efficient Supervised Knowledge Distillation of Vision Transformers. InEuropean Conference on Computer Vision

    2024

  39. [49]

    Zehua Sun, Qiuhong Ke, Hossein Rahmani, Mohammed Bennamoun, Gang Wang, and Jun Liu

  40. [50]

    Zhiyin Shao, Xinyu Zhang, Meng Fang, Zhifeng Lin, Jian Wang, and Changxing Ding. 2022. Learning granularity-unified representations for text-to-image person re-identification. In Pro- ceedings of the 30th acm international conference on multimedia. 5566–5574

    2022

  41. [51]

    Shixiang Tang, Cheng Chen, Qingsong Xie, Meilin Chen, Yizhou Wang, Yuanzheng Ci, Lei Bai, Feng Zhu, Haiyang Yang, Li Yi, et al. 2023. Humanbench: Towards general human-centric perception with projector assisted pretraining. In Proceedings of the IEEE/CVF Conference on Computer...

    2023

  42. [52]

    Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé Jégou. 2021. Training data-efficient image transformers & distillation through attention. In International conference on machine learning. PMLR, 10347–10357

    2021

  43. [53]

    Xuanhan Wang, Xiaojia Chen, Lianli Gao, Jingkuan Song, and Heng Tao Shen. 2024. CPI- Parser: Integrating Causal Properties Into Multiple Human Parsing. IEEE Transactions on Image Processing 33 (2024), 5771–5782

    2024

  44. [54]

    Wei Suo, Mengyang Sun, Kai Niu, Yiqi Gao, Peng Wang, Yanning Zhang, and Qi Wu. 2022. A simple and robust correlation filtering method for text-based person search. In European conference on computer vision. Springer, 726–742. 18

    2022

  45. [55]

    Xuanhan Wang, Jingkuan Song, Xiaojia Chen, Lechao Cheng, Lianli Gao, and Heng Tao Shen

  46. [56]

    Longhui Wei, Shiliang Zhang, Wen Gao, and Qi Tian. 2018. Person transfer gan to bridge domain gap for person re-identification. In Proceedings of the IEEE conference on computer vision and pattern recognition. 79–88

    2018

  47. [57]

    Yufei Xu, Jing Zhang, Qiming Zhang, and Dacheng Tao. 2022. Vitpose: Simple vision transformer baselines for human pose estimation. Advances in neural information processing systems 35 (2022), 38571–38584

    2022

  48. [58]

    Xuanhan Wang, Huimin Deng, Lianli Gao, and Jingkuan Song. 2025. Scale-Aware Pre-Training for Human-Centric Visual Perception: Enabling Lightweight and Generalizable Models. arXiv preprint arXiv:2503.08201 (2025)

    2025 arXiv

  49. [59]

    Hongxu Yin, Pavlo Molchanov, Jose M Alvarez, Zhizhong Li, Arun Mallya, Derek Hoiem, Niraj K Jha, and Jan Kautz. 2020. Dreaming to distill: Data-free knowledge transfer via deepinversion. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 8715–8724

    2020

  50. [60]

    Hang Yu, Yufei Xu, Jing Zhang, Wei Zhao, Ziyu Guan, and Dacheng Tao. 2021. AP-10K: A Benchmark for Animal Pose Estimation in the Wild. In Advances in Neural Information Processing Systems

    2021

  51. [61]

    Junkun Yuan, Xinyu Zhang, Hao Zhou, Jian Wang, Zhongwei Qiu, Zhiyin Shao, Shaofeng Zhang, Sifan Long, Kun Kuang, Kun Yao, et al. 2023. Hap: Structure-aware masked image modeling for human-centric perception. Advances in Neural Information Processing Systems 36 (2023), 50597–50616

    2023

  52. [62]

    Yuhui Yuan, Rao Fu, Lang Huang, Weihong Lin, Chao Zhang, Xilin Chen, and Jingdong Wang. 2021. Hrformer: High-resolution transformer for dense prediction. arXiv preprint arXiv:2110.09408 (2021)

    2021 arXiv

  53. [63]

    Zhendong Yang, Zhe Li, Ailing Zeng, Zexian Li, Chun Yuan, and Yu Li. 2024. ViTKD: Feature-based Knowledge Distillation for Vision Transformers. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). 1379–1388

    2024

  54. [64]

    Liang Zheng, Liyue Shen, Lu Tian, Shengjin Wang, Jingdong Wang, and Qi Tian. 2015. Scalable person re-identification: A benchmark. In Proceedings of the IEEE international conference on computer vision. 1116–1124

    2015

  55. [65]

    Howard Zhong, Samarth Mishra, Donghyun Kim, SouYoung Jin, Rameswar Panda, Hilde Kuehne, Leonid Karlinsky, Venkatesh Saligrama, Aude Oliva, and Rogerio Feris. 2023. Learning human action recognition representations without real humans. Advances in Neural Information Processing ...

    2023

  56. [68]

    Yitian Zhang, Xu Ma, Yue Bai, Huan Wang, and Yun Fu. 2025. Accessing Vision Foun- dation Models via ImageNet-1K. In The Thirteenth International Conference on Learning Representations

    2025

  57. [2020]

    arXiv preprint arXiv:2010.11929 (2020)

    An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

    2010 arXiv

  58. [2021]

    arXiv preprint arXiv:2111.12084 (2021)

    Self-supervised pre-training for transformer-based person re-identification. arXiv preprint arXiv:2111.12084 (2021)

    2021 arXiv

  59. [2022]

    IEEE transactions on pattern analysis and machine intelligence 45, 3 (2022), 3200–3225

    Human action recognition from various data modalities: A review. IEEE transactions on pattern analysis and machine intelligence 45, 3 (2022), 3200–3225

    2022

  60. [2023]

    IEEE Transactions on Cybernetics 53, 11 (2023), 7263–7274

    KE-RCNN: Unifying Knowledge-Based Reasoning Into Part-Level Attribute Parsing. IEEE Transactions on Cybernetics 53, 11 (2023), 7263–7274

    2023

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.