Pith. sign in

REVIEW 2 major objections 5 minor 56 references

Identity-specific motion is learnable from free-throws, but video models skip it for faces and jerseys unless appearance is stripped away.

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 ↗

T0 review · grok-4.5

2026-07-12 01:02 UTC pith:HB3OHGZG

load-bearing objection Clean diagnostic study with a useful new free-throw dataset: models ignore available motion until appearance is stripped; residual shape/pose leakage is real but does not sink the main result. the 2 major comments →

arxiv 2607.03633 v1 pith:HB3OHGZG submitted 2026-07-03 cs.CV

Probing Identity-Specific Motion Signatures: A Controlled Diagnostic Study

Yingtie Lei , Fangxun Liu , Baicheng Wu , Colin Lee , Ziheng Zhang , Junke Yang , Zhiyuan Tao , Xuyan Huang
show 8 more authors
Shuheng Wang William Koran Kyle Park Elijah H Buckwalter Cheng-Hsuan Chiang Tejas Naik Daniel Yi Wei-Lun Chao
This is my paper
classification cs.CV
keywords identity recognitionmotion signaturesvideo modelsappearance shortcutssilhouetteskeletonfree-throwdiagnostic benchmark
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.

This paper asks a diagnostic question: when identity-specific motion is clearly available, do modern video models actually use it to recognize people? The authors build BALLER120, a controlled set of free-throw clips from 120 professional basketball players, plus matched silhouette and skeleton versions that remove faces, jerseys, and texture. With full RGB input, action-recognition backbones reach near-perfect closed-set accuracy, but saliency and appearance-shift tests show they lean on static cues. When the same architectures are trained only on silhouettes or skeletons, they still achieve competitive accuracy, become far more robust when jerseys change, and attend to phase-aligned micro-patterns such as foot placement, elbow trajectory, and torso timing that differ across players. The study therefore establishes that individual motion signatures in a shared skilled action are present and learnable, yet are easily overshadowed by easier appearance shortcuts unless those shortcuts are deliberately removed.

Core claim

Identity-specific motion signatures in free-throw execution are present, informative, and learnable by modern video backbones, but those models preferentially exploit static appearance shortcuts (faces, jersey regions) when both are available; only explicit appearance suppression forces the same architectures onto distinctive, phase-aligned kinematic patterns that remain competitive and more robust under appearance shift.

What carries the argument

BALLER120: a controlled diagnostic dataset of free-throw sequences from 120 NBA players, paired with appearance, silhouette, and skeleton regimes that hold the multi-phase action fixed while suppressing execution-independent cues, allowing direct comparison of what evidence identity classifiers actually use.

Load-bearing premise

That free-throw differences among professional players, after cropping, scale normalization, and mask or skeleton abstraction, mainly reflect stable individual movement habits rather than leftover capture biases still tied to identity.

What would settle it

Train and test the same silhouette or skeleton models after deliberately injecting controlled residual cues (fixed camera-angle clusters per player, or pose-estimator noise correlated with identity) and check whether accuracy and saliency maps still track player-specific execution phases; collapse under those controls would falsify the claim that the recovered signal is pure motion signature.

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

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

2 major / 5 minor

Summary. This paper asks whether modern video models use identity-specific motion when such cues are clearly available, or whether they prefer static appearance shortcuts. It introduces BALLER120, a controlled diagnostic set of 4,583 free-throw clips from 120 NBA players, with paired appearance, silhouette, and skeleton regimes. Using MViTv2, VideoMAEv2, and UniFormerV2 fine-tuned for closed-set identity recognition, the authors show near-ceiling accuracy under a standard split for all regimes (Table 2), sharp collapse of appearance models under an appearance-disjoint split while silhouette/skeleton models remain robust (Table 3), competitive open-set transfer for appearance-suppressed inputs (Fig. 6), limited sensitivity to contour degradation (Table 4), and CAM/pose-region analyses indicating phase-aligned, identity-distinct attention under suppression (Figs. 8–11). Complementary probes (static DINOv3 frames, temporal shuffle, action-recognition control) support a task-induced appearance bias: models exploit the easiest predictive signal unless appearance is suppressed.

Significance. If the result holds, the paper makes a clear diagnostic contribution rather than another unconstrained Re-ID leaderboard entry. BALLER120 is carefully designed to reduce action-level and acquisition confounds while pairing appearance with appearance-suppressed regimes, enabling cue attribution that gait and clothing-change Re-ID datasets do not isolate as cleanly. The multi-backbone, multi-split, ablation, and saliency package is unusually thorough for a diagnostic study and yields a falsifiable, practically relevant message: identity-specific execution cues are learnable and more robust under appearance shift, but standard video fine-tuning will often ignore them. Strengths include explicit multi-regime construction, appearance-disjoint and open-set protocols, contour-degradation and temporal-shuffle controls, and transparent positioning as a probe rather than a general benchmark. These make the work useful for biometrics, sports analytics, and video representation research even if residual non-dynamic cues remain partially entangled.

major comments (2)
  1. [§3–4, Tables 3–5, abstract/§7] The central claim that silhouette/skeleton success reflects identity-specific motion signatures is only partially isolated from residual execution-independent or multi-pose static anatomy. §4’s union-crop + common-scale resize removes absolute height/camera distance, and Table 4/Fig. 7 show contour degradation barely hurts accuracy, but limb proportions, torso aspect, and phase-specific joint configurations (or RTMPose biases correlated with body type) can still identify players without continuous dynamics. Table 5 is the load-bearing control: multi-frame DINOv3 on silhouettes already reaches ~90% Top-1, so ordered video is helpful but not uniquely necessary. Temporal shuffle (Supp. Table F1) and multi-view consistency help, yet the abstract and §7 still state the result primarily as “motion micro-patterns” / “motion signatures.” Please either (i) reframe claims around execution-dependen
  2. [§6.2, Table 2, Supp. Table C1] Skeleton results are reported as supporting appearance-suppressed motion learning, but Table 2 and Supp. Table C1 show strong backbone dependence: only MViTv2 remains near ceiling; VideoMAEv2 collapses (~25% Top-1). The paper notes this and supplies saliency hypotheses in Supp. §C–D, but the main-text claim that “skeleton-only inputs” induce a shift toward motion micro-patterns is overstated relative to the architecture-specific evidence. Either restrict the skeleton claim to MViTv2 throughout the abstract/§6.2, or provide a controlled diagnosis (e.g., denser keypoint rendering, temporal joint features, or architecture ablations) showing when sparse skeletons preserve identity-linked dynamics rather than architecture-specific compatibility.
minor comments (5)
  1. [Figs. 6, 9] Fig. 1 and the teaser caption are effective, but several later figures (e.g., Fig. 9 regional saliency curves, Fig. 6 open-set bars) would benefit from explicit axis units, error bars or seed variance, and a short note on how many clips/identities underlie each bar.
  2. [§5, Eqs. (1)–(3)] The loss combination Lid + Ltri (Eqs. 1–3) is reasonable, but the paper does not report an ablation of classification-only vs. triplet-only training. A short note or table would clarify whether the cue-shift findings depend on the metric-learning term.
  3. [Table 1] Table 1’s “Explain. Support” column is informative but slightly overloaded; a one-sentence definition of what counts as phase-level interpretability support would help readers compare BALLER120 to gait sets.
  4. [§3–4] Minor wording: “execution-independent anatomy” vs. “fixed anatomy” is used somewhat interchangeably in §3–4; aligning terminology would reduce ambiguity with residual shape concerns.
  5. [§6.4, Fig. 12] The action-recognition control (three-pointers vs. free-throws, Fig. 12) is valuable; stating the number of three-point clips and whether the same players/views are used would improve reproducibility.

Circularity Check

0 steps flagged

No significant circularity: empirical diagnostic study whose claims rest on external identity labels, controlled input regimes, and measured accuracy/robustness, not on self-referential definitions or fitted-as-prediction steps.

full rationale

BALLER120 and the reported results form a standard controlled empirical probe rather than a derivation chain. Identity labels are external (NBA player identities); free-throw structure is an observed multi-phase action, not defined from model outputs. Accuracy, mAP, appearance-disjoint robustness (Tables 2–3, Fig. 6), contour-degradation sanity checks (Table 4), temporal-shuffle ablations (Supp. Table F1), and CAM saliency (Figs. 8–11) are measured outcomes under three input regimes constructed by an explicit pipeline (Grounding DINO + SAM2 + RTMPose + union-box resize). No equation equates a fitted parameter to a claimed prediction; no uniqueness theorem or ansatz is imported via self-citation to force the central claim that motion signatures are learnable yet overlooked unless appearance is suppressed. Minor design choices (ball inclusion rule, spatial normalization) and one co-author citation for Finer-CAM are non-load-bearing diagnostics. The paper is therefore self-contained against its own benchmarks; residual shape leakage is a correctness/assumption concern, not circularity.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 2 invented entities

Empirical computer-vision diagnostic paper. Load-bearing content is the dataset construction choices and the assumption that free-throw kinematics carry stable identity signal after appearance suppression. Hyperparameters are standard fine-tuning knobs; no new physical entities are postulated.

free parameters (4)
  • label-smoothing epsilon = 0.1
    Set to 0.1 in the identity classification loss; standard but chosen by authors.
  • triplet margin m = 0.3
    Margin in batch-hard triplet loss; set to 0.3.
  • peak learning rate / weight decay / epochs / batch size = 1e-4 / 0.05 / 100 / 8-16
    AdamW 1e-4, wd 0.05, 100 epochs, batch 8–16; tuned on the standard-split validation.
  • drop-path / dropout rates per backbone = varies (e.g. VideoMAEv2 drop_path 0.0, drop 0.2)
    Slightly modified from original Kinetics configs (Tables A1–A3) after validation on standard split.
axioms (4)
  • domain assumption Professional free-throw execution contains stable, individual-specific kinematic regularities that survive spatial normalization and appearance suppression.
    Core premise of the diagnostic design (§1, §3); without it the silhouette/skeleton accuracy would not be interpretable as motion signatures.
  • domain assumption Frame-wise SAM2 masks and RTMPose skeletons sufficiently remove execution-independent appearance and anatomy while preserving execution-dependent dynamics.
    Stated in §4; residual mask or pose errors could still leak identity.
  • ad hoc to paper CAM-based saliency, when stable within identity, distinct across identities, and phase-aligned, constitutes valid diagnostic evidence of cue use.
    Interpretation criteria defined in §B of the supplement; saliency is treated as support not proof.
  • domain assumption Standard action-recognition pre-training (Kinetics) plus full fine-tuning yields representations that can be fairly compared across appearance/silhouette/skeleton regimes.
    §5; the comparison assumes the backbones are not inherently biased against sparse skeleton input beyond what is measured.
invented entities (2)
  • BALLER120 diagnostic benchmark no independent evidence
    purpose: Controlled multi-regime free-throw dataset that reduces action-level and acquisition biases so identity-specific motion can be isolated and inspected.
    New dataset constructed for this study; no prior public equivalent with paired appearance/silhouette/skeleton free-throw clips of 120 pros.
  • task-induced appearance bias no independent evidence
    purpose: Name for the observed tendency of video models to prefer static appearance for identity tasks but motion for action discrimination.
    Descriptive label introduced in §6.4; not a new physical mechanism, just a behavioral observation.

pith-pipeline@v1.1.0-grok45 · 29796 in / 3062 out tokens · 27768 ms · 2026-07-12T01:02:25.667355+00:00 · methodology

0 comments
read the original abstract

Identity recognition (e.g., person, animal re-identification) has traditionally relied heavily on static appearance cues. Yet motion--consistent, individual-specific dynamics--can provide a complementary and potentially more robust signature, especially when appearance is weak or variable. This raises a fundamental question: when identity-specific motion cues are clearly present, to what extent do modern video models use them for recognition? To investigate this question, we conduct a systematic diagnostic study and introduce BALLER120, a controlled benchmark of 120 professional basketball players performing free-throws. By focusing on the same multi-phase action across individuals, BALLER120 reduces action-level variation and identity-correlated acquisition biases, enabling fine-grained analysis of identity-specific kinematic patterns. We find that modern video models can predict identity accurately from RGB videos, but often rely on static appearance cues such as faces and jersey regions, even when informative motion cues are available. Strikingly, when appearance is suppressed through silhouette-only or skeleton-only inputs, the same model architectures shift toward motion micro-patterns (e.g., foot placement and elbow bending). Despite containing less visual information, appearance-suppressed representations achieve competitive accuracy and stronger robustness to appearance shifts. Our qualitative analyses further show that appearance-suppressed models attend to distinctive motion patterns across individuals. Overall, our study demonstrates that identity-specific motion signatures are present, informative, and learnable, but modern video models may overlook them in favor of easier static shortcuts unless appearance cues are explicitly suppressed.

Figures

Figures reproduced from arXiv: 2607.03633 by Baicheng Wu, Cheng-Hsuan Chiang, Colin Lee, Daniel Yi, Elijah H Buckwalter, Fangxun Liu, Junke Yang, Kyle Park, Shuheng Wang, Tejas Naik, Wei-Lun Chao, William Koran, Xuyan Huang, Yingtie Lei, Zhiyuan Tao, Ziheng Zhang.

Figure 1
Figure 1. Figure 1: Overview of our diagnostic probe for identity-specific motion signatures. Given the same free-throw sequence, models can rely [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Gait benchmarks vs. BALLER120. Each row shows a different individual. BALLER120 is designed as a diagnostic probe: it focuses on free-throws, a multi-phase skilled action that naturally aligns execution across individuals, making it easier to visualize and verify the fine-grained identity-specific motion cues used by a model. In contrast, CASIA-B [16] focuses on identity recognition from cyclic walking mot… view at source ↗
Figure 3
Figure 3. Figure 3: Data pipeline. (a) We localize all players in the first [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Identity probing with action-recognition backbones. Each video, in one of three input regimes (appearance, silhouette, or skeleton), is passed through a fully fine-tuned ( ) backbone, whose embedding is optimized with an identity classification loss Lid and a triplet loss Ltri. 5. Repurposing Action-Recognition Backbones for Identity Probing Action recognition (AR) and identity recognition pursue different… view at source ↗
Figure 4
Figure 4. Figure 4: Per-identity clip distribution across the 120 NBA play [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Open-set performance under the appearance-disjoint set [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Saliency maps for the same player (Al Horford). The [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Pose-based quantification of attention dynamics. [PITH_FULL_IMAGE:figures/full_fig_p007_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Saliency maps of the same player across different cam [PITH_FULL_IMAGE:figures/full_fig_p008_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Saliency maps for Ja Morant and Al Horford perform [PITH_FULL_IMAGE:figures/full_fig_p008_11.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

56 extracted references · 8 linked inside Pith

  1. [1]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 1

    2009

  2. [2]

    Imagenet classification with deep convolutional neural net- works.Advances in neural information processing systems, 25, 2012

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural net- works.Advances in neural information processing systems, 25, 2012

    2012

  3. [3]

    Deep learning for person re- identification: A survey and outlook.IEEE transactions on pattern analysis and machine intelligence, 44(6):2872–2893, 2021

    Mang Ye, Jianbing Shen, Gaojie Lin, Tao Xiang, Ling Shao, and Steven CH Hoi. Deep learning for person re- identification: A survey and outlook.IEEE transactions on pattern analysis and machine intelligence, 44(6):2872–2893, 2021

    2021

  4. [4]

    Fine-grained image analysis with deep learning: A survey

    Xiu-Shen Wei, Yi-Zhe Song, Oisin Mac Aodha, Jianxin Wu, Yuxin Peng, Jinhui Tang, Jian Yang, and Serge Belongie. Fine-grained image analysis with deep learning: A survey. IEEE transactions on pattern analysis and machine intelli- gence, 44(12):8927–8948, 2021. 1

    2021

  5. [5]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. InInternational conference on machine learning, pages 8748–8763. PmLR, 2021. 2

    2021

  6. [6]

    A convnet for the 2020s

    Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feicht- enhofer, Trevor Darrell, and Saining Xie. A convnet for the 2020s. InProceedings of the IEEE/CVF conference on com- puter vision and pattern recognition, pages 11976–11986, 2022

    2022

  7. [7]

    Sigmoid loss for language image pre-training

    Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. Sigmoid loss for language image pre-training. InProceedings of the IEEE/CVF international conference on computer vision, pages 11975–11986, 2023

    2023

  8. [8]

    Internvideo2.5: Empowering video mllms with long and rich context modeling.arXiv preprint arXiv:2501.12386, 2025

    Yi Wang, Xinhao Li, Ziang Yan, Yinan He, Jiashuo Yu, Xi- angyu Zeng, Chenting Wang, Changlian Ma, Haian Huang, Jianfei Gao, et al. Internvideo2.5: Empowering video mllms with long and rich context modeling.arXiv preprint arXiv:2501.12386, 2025

    Pith/arXiv arXiv 2025

  9. [9]

    Dinov3.arXiv preprint arXiv:2508.10104, 2025

    Oriane Sim ´eoni, Huy V V o, Maximilian Seitzer, Federico Baldassarre, Maxime Oquab, Cijo Jose, Vasil Khalidov, Marc Szafraniec, Seungeun Yi, Micha ¨el Ramamonjisoa, et al. Dinov3.arXiv preprint arXiv:2508.10104, 2025. 2, 8

    Pith/arXiv arXiv 2025

  10. [10]

    Passos, Rafael Gonc ¸alves Pires, Daniel Felipe Silva Santos, Lucas Pascotti Valem, Thierry P

    Claudio Filipi Gonc ¸alves dos Santos, Diego De Souza Oliveira, Leandro A. Passos, Rafael Gonc ¸alves Pires, Daniel Felipe Silva Santos, Lucas Pascotti Valem, Thierry P. Moreira, Marcos Cleison S. Santana, Mateus Roder, Jo Paulo Papa, et al. Gait recognition based on deep learning: A survey.ACM Computing Surveys (CSUR), 55(2):1–34, 2022. 2

    2022

  11. [11]

    A comprehensive survey on deep gait recog- nition: Algorithms, datasets, and challenges.IEEE Trans- actions on Biometrics, Behavior, and Identity Science, 7(2): 270–292, 2024

    Chuanfu Shen, Shiqi Yu, Jilong Wang, George Q Huang, and Liang Wang. A comprehensive survey on deep gait recog- nition: Algorithms, datasets, and challenges.IEEE Trans- actions on Biometrics, Behavior, and Identity Science, 7(2): 270–292, 2024. 2

    2024

  12. [12]

    Mvitv2: Improved multiscale vision transform- ers for classification and detection

    Yanghao Li, Chao-Yuan Wu, Haoqi Fan, Karttikeya Man- galam, Bo Xiong, Jitendra Malik, and Christoph Feicht- enhofer. Mvitv2: Improved multiscale vision transform- ers for classification and detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4804–4814, 2022. 2, 6, 7, 8

    2022

  13. [13]

    Videomae v2: Scaling video masked autoencoders with dual masking

    Limin Wang, Bingkun Huang, Zhiyu Zhao, Zhan Tong, Yi- nan He, Yi Wang, Yali Wang, and Yu Qiao. Videomae v2: Scaling video masked autoencoders with dual masking. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 14549–14560, 2023. 2, 6, 7, 8, 5

    2023

  14. [14]

    Uniformerv2: Unlocking the po- tential of image vits for video understanding

    Kunchang Li, Yali Wang, Yinan He, Yizhuo Li, Yi Wang, Limin Wang, and Yu Qiao. Uniformerv2: Unlocking the po- tential of image vits for video understanding. InProceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 1632–1643, 2023. 2, 5, 6, 7, 8

    2023

  15. [15]

    Learning deep features for discrimina- tive localization

    Bolei Zhou, Aditya Khosla, Agata Lapedriza, Aude Oliva, and Antonio Torralba. Learning deep features for discrimina- tive localization. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 2921–2929,

  16. [16]

    A framework for evaluating the effect of view angle, clothing and carrying condition on gait recognition

    Shiqi Yu, Daoliang Tan, and Tieniu Tan. A framework for evaluating the effect of view angle, clothing and carrying condition on gait recognition. In18th international con- ference on pattern recognition (ICPR’06), pages 441–444. IEEE, 2006. 3, 5

    2006

  17. [17]

    Person re-identification in the wild

    Liang Zheng, Hengheng Zhang, Shaoyan Sun, Manmohan Chandraker, Yi Yang, and Qi Tian. Person re-identification in the wild. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 1367–1376, 2017. 2

    2017

  18. [18]

    Past, present and future approaches us- ing computer vision for animal re-identification from cam- era trap data.Methods in Ecology and Evolution, 10(4):461– 470, 2019

    Stefan Schneider, Graham W Taylor, Stefan Linquist, and Stefan C Kremer. Past, present and future approaches us- ing computer vision for animal re-identification from cam- era trap data.Methods in Ecology and Evolution, 10(4):461– 470, 2019

    2019

  19. [19]

    Gait recognition in the wild: A benchmark

    Zheng Zhu, Xianda Guo, Tian Yang, Junjie Huang, Jiankang Deng, Guan Huang, Dalong Du, Jiwen Lu, and Jie Zhou. Gait recognition in the wild: A benchmark. InProceedings of the IEEE/CVF international conference on computer vi- sion, pages 14789–14799, 2021. 3, 5

    2021

  20. [20]

    Gait recognition in the wild with dense 3d representations and a benchmark

    Jinkai Zheng, Xinchen Liu, Wu Liu, Lingxiao He, Cheng- gang Yan, and Tao Mei. Gait recognition in the wild with dense 3d representations and a benchmark. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 20228–20237, 2022. 3, 5

    2022

  21. [21]

    An in-depth ex- ploration of person re-identification and gait recognition in cloth-changing conditions

    Weijia Li, Saihui Hou, Chunjie Zhang, Chunshui Cao, Xu Liu, Yongzhen Huang, and Yao Zhao. An in-depth ex- ploration of person re-identification and gait recognition in cloth-changing conditions. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13824–13833, 2023. 2, 3, 5

    2023

  22. [22]

    Deep- reid: Deep filter pairing neural network for person re- identification

    Wei Li, Rui Zhao, Tong Xiao, and Xiaogang Wang. Deep- reid: Deep filter pairing neural network for person re- identification. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 152–159,

  23. [23]

    Scalable person re-identification: A benchmark

    Liang Zheng, Liyue Shen, Lu Tian, Shengjin Wang, Jing- dong Wang, and Qi Tian. Scalable person re-identification: A benchmark. InProceedings of the IEEE international con- ference on computer vision, pages 1116–1124, 2015. 3, 5

    2015

  24. [24]

    Person transfer gan to bridge domain gap for person re- identification

    Longhui Wei, Shiliang Zhang, Wen Gao, and Qi Tian. Person transfer gan to bridge domain gap for person re- identification. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 79–88, 2018. 3, 5

    2018

  25. [25]

    Person re-identification by video ranking

    Taiqing Wang, Shaogang Gong, Xiatian Zhu, and Shengjin Wang. Person re-identification by video ranking. In European conference on computer vision, pages 688–703. Springer, 2014. 3, 5

    2014

  26. [26]

    Mars: A video benchmark for large-scale person re-identification

    Liang Zheng, Zhi Bie, Yifan Sun, Jingdong Wang, Chi Su, Shengjin Wang, and Qi Tian. Mars: A video benchmark for large-scale person re-identification. InEuropean conference on computer vision, pages 868–884. Springer, 2016. 3, 5

    2016

  27. [27]

    Person re- identification by contour sketch under moderate clothing change.IEEE transactions on pattern analysis and machine intelligence, 43(6):2029–2046, 2019

    Qize Yang, Ancong Wu, and Wei-Shi Zheng. Person re- identification by contour sketch under moderate clothing change.IEEE transactions on pattern analysis and machine intelligence, 43(6):2029–2046, 2019. 3, 5

    2029

  28. [28]

    Long-term cloth-changing person re-identification

    Xuelin Qian, Wenxuan Wang, Li Zhang, Fangrui Zhu, Yan- wei Fu, Tao Xiang, Yu-Gang Jiang, and Xiangyang Xue. Long-term cloth-changing person re-identification. InPro- ceedings of the Asian conference on computer vision, 2020. 3, 5

    2020

  29. [29]

    Deepchange: A long-term per- son re-identification benchmark with clothes change

    Peng Xu and Xiatian Zhu. Deepchange: A long-term per- son re-identification benchmark with clothes change. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11196–11205, 2023. 3, 5

    2023

  30. [30]

    Clothes-changing person re-identification with rgb modality only

    Xinqian Gu, Hong Chang, Bingpeng Ma, Shutao Bai, Shiguang Shan, and Xilin Chen. Clothes-changing person re-identification with rgb modality only. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1060–1069, 2022. 3, 5

    2022

  31. [31]

    Mevid: Multi-view extended videos with identities for video person re-identification

    Daniel Davila, Dawei Du, Bryon Lewis, Christopher Funk, Joseph Van Pelt, Roderic Collins, Kellie Corona, Matt Brown, Scott McCloskey, Anthony Hoogs, and Brian Clipp. Mevid: Multi-view extended videos with identities for video person re-identification. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 1634–164...

    2023

  32. [32]

    Wildlifereid-10k: Wildlife re-identification dataset with 10k individual animals

    Lukas Adam, V ojtech Cerm ´ak, Kostas Papafitsoros, and Lukas Picek. Wildlifereid-10k: Wildlife re-identification dataset with 10k individual animals. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2124–2134, 2025. 3

    2025

  33. [33]

    Understanding the recognition of facial identity and facial expression.Nature Reviews Neuroscience, 6(8):641–651, 2005

    Andrew J Calder and Andrew W Young. Understanding the recognition of facial identity and facial expression.Nature Reviews Neuroscience, 6(8):641–651, 2005. 3

    2005

  34. [34]

    How iris recognition works

    John Daugman. How iris recognition works. InThe essential guide to image processing, pages 715–739. Elsevier, 2009

    2009

  35. [35]

    On the individuality of fingerprints.IEEE Transactions on pattern analysis and machine intelligence, 24(8):1010–1025, 2002

    Sharath Pankanti, Salil Prabhakar, and Anil K Jain. On the individuality of fingerprints.IEEE Transactions on pattern analysis and machine intelligence, 24(8):1010–1025, 2002

    2002

  36. [36]

    Personal authentication using palm-print fea- tures.Pattern recognition, 36(2):371–381, 2003

    Chin-Chuan Han, Hsu-Liang Cheng, Chih-Lung Lin, and Kuo-Chin Fan. Personal authentication using palm-print fea- tures.Pattern recognition, 36(2):371–381, 2003. 3

    2003

  37. [37]

    Dygait: Exploiting dynamic representations for high-performance gait recogni- tion

    Ming Wang, Xianda Guo, Beibei Lin, Tian Yang, Zheng Zhu, Lincheng Li, Shunli Zhang, and Xin Yu. Dygait: Exploiting dynamic representations for high-performance gait recogni- tion. InProceedings of the IEEE/CVF international confer- ence on computer vision, pages 13424–13433, 2023. 3

    2023

  38. [38]

    Biggait: Learning gait representation you want by large vision models

    Dingqiang Ye, Chao Fan, Jingzhe Ma, Xiaoming Liu, and Shiqi Yu. Biggait: Learning gait representation you want by large vision models. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 200–210, 2024. 3, 6

    2024

  39. [39]

    The kinetics hu- man action video dataset.arXiv preprint arXiv:1705.06950,

    Will Kay, Joao Carreira, Karen Simonyan, Brian Zhang, Chloe Hillier, Sudheendra Vijayanarasimhan, Fabio Viola, Tim Green, Trevor Back, Paul Natsev, et al. The kinetics hu- man action video dataset.arXiv preprint arXiv:1705.06950,

    Pith/arXiv arXiv

  40. [40]

    A short note about kinetics- 600.arXiv preprint arXiv:1808.01340, 2018

    Joao Carreira, Eric Noland, Andras Banki-Horvath, Chloe Hillier, and Andrew Zisserman. A short note about kinetics- 600.arXiv preprint arXiv:1808.01340, 2018

    Pith/arXiv arXiv 2018

  41. [41]

    A short note on the kinetics-700 human action dataset.arXiv preprint arXiv:1907.06987, 2019

    Joao Carreira, Eric Noland, Chloe Hillier, and Andrew Zis- serman. A short note on the kinetics-700 human action dataset.arXiv preprint arXiv:1907.06987, 2019. 3

    Pith/arXiv arXiv 1907

  42. [42]

    Moments in time dataset: one million videos for event understanding

    Mathew Monfort, Alex Andonian, Bolei Zhou, Kandan Ra- makrishnan, Sarah Adel Bargal, Tom Yan, Lisa Brown, Quanfu Fan, Dan Gutfreund, Carl V ondrick, et al. Moments in time dataset: one million videos for event understanding. IEEE transactions on pattern analysis and machine intelli- gence, 42(2):502–508, 2019. 3

    2019

  43. [43]

    The” something something” video database for learning and evaluating visual common sense

    Raghav Goyal, Samira Ebrahimi Kahou, Vincent Michal- ski, Joanna Materzynska, Susanne Westphal, Heuna Kim, Valentin Haenel, Ingo Fruend, Peter Yianilos, Moritz Mueller-Freitag, et al. The” something something” video database for learning and evaluating visual common sense. InProceedings of the IEEE international conference on com- puter vision, pages 5842...

    2017

  44. [44]

    Grounding dino: Marrying dino with grounded pre-training for open-set object detection

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Qing Jiang, Chunyuan Li, Jianwei Yang, Hang Su, et al. Grounding dino: Marrying dino with grounded pre-training for open-set object detection. InEuro- pean conference on computer vision, pages 38–55. Springer,

  45. [45]

    Sam 2: Segment anything in images and videos

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman R¨adle, Chloe Rolland, Laura Gustafson, Eric Mintun, Junt- ing Pan, Kalyan Vasudev Alwala, Nicolas Carion, Chao- Yuan Wu, Ross Girshick, Piotr Doll´ar, and Christoph Feicht- enhofer. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:...

    Pith/arXiv arXiv 2024

  46. [46]

    Rtmpose: Real- time multi-person pose estimation based on mmpose.arXiv preprint arXiv:2303.07399, 2023

    Tao Jiang, Peng Lu, Li Zhang, Ningsheng Ma, Rui Han, Chengqi Lyu, Yining Li, and Kai Chen. Rtmpose: Real- time multi-person pose estimation based on mmpose.arXiv preprint arXiv:2303.07399, 2023. 4

    Pith/arXiv arXiv 2023

  47. [47]

    Ex- ploring deep models for practical gait recognition.arXiv preprint arXiv:2303.03301, 2023

    Chao Fan, Saihui Hou, Yongzhen Huang, and Shiqi Yu. Ex- ploring deep models for practical gait recognition.arXiv preprint arXiv:2303.03301, 2023. 6

    Pith/arXiv arXiv 2023

  48. [48]

    Decoupled weight decay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. InICLR, 2019. 6

    2019

  49. [49]

    ” why should i trust you?” explaining the predictions of any classifier

    Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. ” why should i trust you?” explaining the predictions of any classifier. InProceedings of the 22nd ACM SIGKDD interna- tional conference on knowledge discovery and data mining, pages 1135–1144, 2016. 2

    2016

  50. [50]

    A unified approach to interpreting model predictions.Advances in neural informa- tion processing systems, 30, 2017

    Scott M Lundberg and Su-In Lee. A unified approach to interpreting model predictions.Advances in neural informa- tion processing systems, 30, 2017. 2

    2017

  51. [51]

    How to explain individual classification decisions.The Jour- nal of Machine Learning Research, 11:1803–1831, 2010

    David Baehrens, Timon Schroeter, Stefan Harmeling, Mo- toaki Kawanabe, Katja Hansen, and Klaus-Robert M ¨uller. How to explain individual classification decisions.The Jour- nal of Machine Learning Research, 11:1803–1831, 2010. 2

    2010

  52. [52]

    Grad-cam: Visual explanations from deep networks via gradient-based localization

    Ramprasaath R Selvaraju, Michael Cogswell, Abhishek Das, Ramakrishna Vedantam, Devi Parikh, and Dhruv Batra. Grad-cam: Visual explanations from deep networks via gradient-based localization. InProceedings of the IEEE in- ternational conference on computer vision, pages 618–626,

  53. [53]

    Grad-cam++: General- ized gradient-based visual explanations for deep convolu- tional networks

    Aditya Chattopadhay, Anirban Sarkar, Prantik Howlader, and Vineeth N Balasubramanian. Grad-cam++: General- ized gradient-based visual explanations for deep convolu- tional networks. In2018 IEEE winter conference on appli- cations of computer vision (WACV), pages 839–847. IEEE, 2018

    2018

  54. [54]

    Score-cam: Score-weighted visual explanations for convolutional neural networks

    Haofan Wang, Zifan Wang, Mengnan Du, Fan Yang, Zijian Zhang, Sirui Ding, Piotr Mardziel, and Xia Hu. Score-cam: Score-weighted visual explanations for convolutional neural networks. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 24–25, 2020. 2

    2020

  55. [55]

    Finer-cam: Spotting the difference reveals finer details for visual explanation

    Ziheng Zhang, Jianyang Gu, Arpita Chowdhury, Zheda Mai, David Carlyn, Tanya Berger-Wolf, Yu Su, and Wei-Lun Chao. Finer-cam: Spotting the difference reveals finer details for visual explanation. InProceedings of the Computer Vi- sion and Pattern Recognition Conference, pages 9611–9620,

  56. [56]

    pure motion

    2 Probing Identity-Specific Motion Signatures: A Controlled Diagnostic Study Supplementary Material This supplementary material provides additional imple- mentation details and analyses that support the diagnostic findings in the main paper. It is organized as follows: • Appendix A reports the backbone configurations and training hyperparameters used in o...

    2097