Pith. sign in

REVIEW 3 major objections 2 minor 50 references

GaitCrafter: Diffusion Model for Biometric Preserving Gait Synthesis

T0 review · 3 major / 2 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read GaitCrafter trains a video diffusion model from scratch on silhouette data to generate temporally consistent, identity-preserving gait sequences, and shows that adding these synthetic samples improves gait recognition, especially under…

desk verdict A plausible gait-diffusion abstract attached to an unrelated condensed-matter full text; no actual method to review. read the letter →

arxiv 2508.13300 v1 pith:QR6TDNII submitted 2025-08-18 cs.CV cs.AI

Sirshapan Mitra , Yogesh S. Rawat This is my paper · ORCID
classification cs.CVcs.AI
keywords gaitrecognitiondiffusionmodelsilhouettesynthesisidentitypreservationprivacy-preservinggenerationvideodataaugmentationcontrollable
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

GaitCrafter tries to solve the data scarcity and privacy problem in gait recognition by synthesizing realistic gait sequences directly in the silhouette domain using a video diffusion model trained from scratch. The paper claims the model produces temporally consistent, identity-preserving sequences and that incorporating these synthetic samples into the recognition pipeline improves performance, particularly under challenging conditions. It also introduces a way to generate novel identities not present in the training data by interpolating identity embeddings, which preserves the privacy of real subjects while providing useful training data. If these claims hold, gait recognition can scale without collecting more labeled real-world gait data and without exposing identities.

What carries the argument

The central object is the video diffusion model operating on gait silhouette sequences, trained from scratch on silhouette data alone. Its role is to model the joint distribution of silhouette frames so that sampled sequences are temporally coherent and identity-specific. The identity-preservation and novel-identity mechanism is carried by the identity embedding space: interpolating between identity embeddings produces new identities with plausible gait patterns, while controlling covariates is achieved by conditioning the diffusion process on input attributes like clothing, carried objects, and view angle.

What would settle it

Run a standard gait recognition benchmark where a recognition model is trained on real data augmented with GaitCrafter synthetic sequences and compared against the same model trained only on real data; if performance does not improve, especially on the challenging conditions highlighted in the paper, the central claim fails. Additionally, generate novel identities by interpolating identity embeddings and test whether those identities are recognized as distinct and consistent across views; if they are confused with their parent identities or lack temporal consistency, the novel-identity mechanism is not sound.

Watch

Extended reading notes

Core claim

The central claim is that a video diffusion model, trained exclusively on gait silhouette data rather than on simulated environments or alternative generative models, can synthesize realistic gait sequences that are temporally consistent and preserve identity. The generation is controllable through conditioning on covariates such as clothing, carried objects, and view angle. The authors further discover that interpolating identity embeddings yields novel, synthetic individuals with unique and consistent gait patterns, and that these synthetic sequences, when added to the training set, improve gait recognition performance, especially under difficult conditions.

Load-bearing premise

The load-bearing premise is that the identity embedding space learned from real subjects is smooth enough that linear interpolation produces valid, recognizable new identities, and that synthetic silhouette sequences transfer to real-world recognition.

Editorial extensions

If this is right

  • Gait recognition systems can be augmented with large volumes of synthetic silhouettes, reducing the need for extensive real data collection and labeling.
  • Privacy is strengthened because novel identities generated by embedding interpolation are not direct copies of real subjects, allowing training on synthetic gait data without exposing original identities.
  • Controllable generation of gait sequences across covariates makes it possible to systematically test and improve recognition robustness to variations in clothing, carried objects, and viewpoint.
  • The approach could be extended to other silhouette-based biometric or action recognition tasks where data collection is expensive or privacy-sensitive.
  • Interpolated identity embeddings suggest a continuous identity manifold that may support applications like gait editing or virtual avatar animation.

Reading between the lines

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

  • The success of the novel-identity mechanism hinges on the identity embedding space being smooth and semantically meaningful; if interpolation produces non-physical or ambiguous gaits, the privacy-preserving benefit would weaken, and this assumption is not directly proven in the abstract.
  • A likely practical test is whether models trained on synthetic silhouettes transfer to real-world recognition when the training and test domains differ in background, resolution, or camera setup, since the silhouette domain may not fully capture real-world variation.
  • The claim that synthetic data improves recognition could be extended by measuring the marginal benefit of each covariate condition, which would tell practitioners which synthetic variations matter most for robustness.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 2 minor

Summary. The submission presents an abstract for 'GaitCrafter,' a proposed diffusion-based framework for synthesizing gait silhouette sequences with identity preservation, controllability over covariates, and improved gait recognition via synthetic data, including generation of novel identities by interpolating identity embeddings. The full text supplied, however, is an unrelated condensed-matter manuscript on charge density wave fluctuations in hole-doped kagome metals, containing no material on gait recognition, diffusion models, or silhouettes. The referee therefore cannot verify any of the abstract's claims against a described method, experiments, or evaluation.

Significance. If the claims in the abstract were supported by a proper manuscript, the contribution could be significant: privacy-preserving synthetic gait data that improves recognition would address a real bottleneck in biometrics. The proposed mechanism of generating novel identities by interpolating identity embeddings is potentially interesting, though its validity depends on the smoothness of the embedding space and on transfer of synthetic silhouettes to real data, neither of which can be assessed from the submitted text. As submitted, the significance cannot be evaluated because the method and results are absent.

major comments (3)
  1. [Full Text] The full text (arXiv:2508.13290v1, 'Persistence of charge density wave fluctuations...') is a condensed-matter physics paper by Kongruengkit et al. on CsV3Sb5−xSnx; it contains no description of GaitCrafter, no gait dataset, no diffusion model, and no recognition experiments. This is a load-bearing omission: none of the abstract's central claims can be checked against the submitted manuscript, and the submission is not the paper described by its own abstract.
  2. [Abstract] The abstract states that incorporating synthetic samples 'leads to improved performance, especially under challenging conditions,' but it reports no quantitative results, no baselines, no evaluation protocol, and no datasets. Even if the correct full text were provided, the abstract alone provides insufficient evidence for this performance claim; concrete numbers, comparison methods, and statistical significance are needed.
  3. [Abstract] The abstract's proposal to generate novel identities by interpolating identity embeddings raises a potential circularity concern: if the identity embeddings are produced by the same recognition model that is later trained on the synthetic data, gains measured in that embedding space could reflect self-consistency rather than improved biometric discrimination. The manuscript needs to specify the source of embeddings and the evaluation protocol to rule this out; the submitted text does not address it.
minor comments (2)
  1. [Full Text] The arXiv identifier in the full-text header (2508.13290) differs from the identifier in the assignment (2508.13300); please confirm that the correct manuscript was uploaded.
  2. [Abstract] The abstract mentions 'controllable' generation conditioned on clothing, carried objects, and view angle, but provides no details on how these covariates are encoded or disentangled; a proper manuscript would need to describe the conditioning mechanism.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; the supplied full text is a different paper, so the GaitCrafter derivation chain cannot be checked for circularity.

full rationale

The submitted 'FULL TEXT' is arXiv:2508.13290v1, 'Persistence of charge density wave fluctuations in the absence of long-range order in a hole-doped kagome metal' by Kongruengkit, Capa Salinas, Pokharel, Ortiz, Wilson, and Harter. This is a condensed-matter physics paper about CDW fluctuations in AV3Sb5, with no mention of gait recognition, silhouettes, diffusion models, or identity embeddings. The abstract's claims about GaitCrafter—training a video diffusion model from scratch on gait silhouette data, generating temporally consistent and identity-preserving sequences, interpolating identity embeddings to create novel identities, and improving recognition performance—are therefore not supported by any equations, architecture details, training procedures, or evaluation protocols in the supplied text. Because the claimed derivation chain is absent, no specific reduction can be exhibited, and no circularity step can be identified under the requirement that circularity be demonstrated by quoting the paper and showing the reduction. I also examined the supplied CDW text on its own terms. Its extrapolated quantum phase transition at x* is obtained from a quadratic fit to magnetization data from Ref. 16 and then checked against an independent transient-reflectivity observable, so the prediction is not constructed from its own confirmatory data. The alpha-mode intensity is used as a symmetry-breaking probe via an external mode assignment from Ref. 11, not as a definitional restatement of the conclusion that CDW fluctuations persist. Thus there is no self-definitional step, no fitted input renamed as a prediction, and no load-bearing self-citation chain. The mismatch between the abstract and the full text is a serious factual and evidentiary problem for the submission, but it is not a circularity, and it does not raise the circularity score.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

Abstract only; no explicit free parameters are visible. The central claims rely on the validity of two domain assumptions: the smoothness/interpolatability of the identity embedding space and the transferability of synthetic silhouette data to real gait recognition. Neither is verifiable from the abstract.

assumptions (2)
  • domain assumption Identity embeddings are linearly interpolatable to produce valid, recognizable novel identities.
    The abstract's mechanism of novel identity generation assumes the embedding space is smooth and semantically linear.
  • domain assumption Synthetic silhouette-domain gait sequences transfer to real-world gait recognition.
    The claimed recognition gains from synthetic samples depend on domain consistency between generated silhouettes and real data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GaitCrafter: Diffusion Model for Biometric Preserving Gait Synthesis." pith.science (2026). https://pith.science/paper/QR6TDNII

@misc{pith2026250813300,
  author       = {Pith},
  title        = {Pith review of: GaitCrafter: Diffusion Model for Biometric Preserving Gait Synthesis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QR6TDNII}},
  note         = {Machine review of arXiv:2508.13300}
}
read the original abstract

Gait recognition is a valuable biometric task that enables the identification of individuals from a distance based on their walking patterns. However, it remains limited by the lack of large-scale labeled datasets and the difficulty of collecting diverse gait samples for each individual while preserving privacy. To address these challenges, we propose GaitCrafter, a diffusion-based framework for synthesizing realistic gait sequences in the silhouette domain. Unlike prior works that rely on simulated environments or alternative generative models, GaitCrafter trains a video diffusion model from scratch, exclusively on gait silhouette data. Our approach enables the generation of temporally consistent and identity-preserving gait sequences. Moreover, the generation process is controllable-allowing conditioning on various covariates such as clothing, carried objects, and view angle. We show that incorporating synthetic samples generated by GaitCrafter into the gait recognition pipeline leads to improved performance, especially under challenging conditions. Additionally, we introduce a mechanism to generate novel identities-synthetic individuals not present in the original dataset-by interpolating identity embeddings. These novel identities exhibit unique, consistent gait patterns and are useful for training models while maintaining privacy of real subjects. Overall, our work takes an important step toward leveraging diffusion models for high-quality, controllable, and privacy-aware gait data generation.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 38 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Activity-biometrics: Person identification from daily activities

    Shehreen Azad and Yogesh Singh Rawat. Activity-biometrics: Person identification from daily activities. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 287--296, 2024

    work page 2024

  3. [3]

    Disenq: Disentangling q-former for activity-biometrics

    Shehreen Azad and Yogesh S Rawat. Disenq: Disentangling q-former for activity-biometrics. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025

    work page 2025

  4. [4]

    Stable video diffusion: Scaling latent video diffusion models to large datasets

    Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram Voleti, Adam Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets. arXiv preprint arXiv:2311.15127, 2023

    arXiv 2023

  5. [5]

    Lagrange motion analysis and view embeddings for improved gait recognition

    Tianrui Chai, Annan Li, Shaoxiong Zhang, Zilong Li, and Yunhong Wang. Lagrange motion analysis and view embeddings for improved gait recognition. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 20217--20226, 2022

    work page 2022

  6. [6]

    Gaitset: Regarding gait as a set for cross-view gait recognition

    Hanqing Chao, Yiwei He, Junping Zhang, and Jianfeng Feng. Gaitset: Regarding gait as a set for cross-view gait recognition. In AAAI Conference on Artificial Intelligence, 2018

    work page 2018

  7. [7]

    Videocrafter2: Overcoming data limitations for high-quality video diffusion models

    Haoxin Chen, Yong Zhang, Xiaodong Cun, Menghan Xia, Xintao Wang, Chao Weng, and Ying Shan. Videocrafter2: Overcoming data limitations for high-quality video diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7310--7320, 2024

    work page 2024

  8. [8]

    Gabriellav2: Towards better generalization in surveillance videos for action detection

    Ishan Dave, Zacchaeus Scheffer, Akash Kumar, Sarah Shiraz, Yogesh Singh Rawat, and Mubarak Shah. Gabriellav2: Towards better generalization in surveillance videos for action detection. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, pages 122--132, 2022

    work page 2022

Show all 50 references
  1. [9]

    Versatilegait: a large-scale synthetic gait dataset with fine-grainedattributes and complicated scenarios

    Huanzhang Dou, Wenhu Zhang, Pengyi Zhang, Yuhan Zhao, Songyuan Li, Zequn Qin, Fei Wu, Lin Dong, and Xi Li. Versatilegait: a large-scale synthetic gait dataset with fine-grainedattributes and complicated scenarios. arXiv preprint arXiv:2101.01394, 2021

    2021 arXiv

  2. [10]

    Metagait: Learning to learn an omni sample adaptive representation for gait recognition

    Huanzhang Dou, Pengyi Zhang, Wei Su, Yunlong Yu, and Xi Li. Metagait: Learning to learn an omni sample adaptive representation for gait recognition. ArXiv, abs/2306.03445, 2023

    2023 arXiv

  3. [11]

    Gaitpart: Temporal part-based model for gait recognition

    Chao Fan, Yunjie Peng, Chunshui Cao, Xu Liu, Saihui Hou, Jiannan Chi, Yongzhen Huang, Qing Li, and Zhiqiang He. Gaitpart: Temporal part-based model for gait recognition. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14213--14221, 2020

    2020

  4. [12]

    Exploring deep models for practical gait recognition

    Chao Fan, Saihui Hou, Yongzhen Huang, and Shiqi Yu. Exploring deep models for practical gait recognition. arXiv preprint arXiv:2303.03301, 2023 a

    2023 arXiv

  5. [13]

    Learning gait representation from massive unlabelled walking videos: A benchmark

    Chao Fan, Saihui Hou, Jilong Wang, Yongzhen Huang, and Shiqi Yu. Learning gait representation from massive unlabelled walking videos: A benchmark. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45 0 (12): 0 14920--14937, 2023 b

    2023

  6. [14]

    Opengait: Revisiting gait recognition towards better practicality

    Chao Fan, Junhao Liang, Chuanfu Shen, Saihui Hou, Yongzhen Huang, and Shiqi Yu. Opengait: Revisiting gait recognition towards better practicality. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9707--9716, 2023 c

    2023

  7. [15]

    Stpro: Spatial and temporal progressive learning for weakly supervised spatio-temporal grounding

    Aaryan Garg, Akash Kumar, and Yogesh S Rawat. Stpro: Spatial and temporal progressive learning for weakly supervised spatio-temporal grounding. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 3384--3394, 2025

    2025

  8. [16]

    Zhang, Shaoqing Ren, and Jian Sun

    Kaiming He, X. Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770--778, 2015

    2016

  9. [17]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. Advances in neural information processing systems, 33: 0 6840--6851, 2020

    2020

  10. [18]

    Jonathan Ho, Tim Salimans, Alexey Gritsenko, William Chan, Mohammad Norouzi, and David J. Fleet. Video diffusion models. ArXiv, abs/2204.03458, 2022

    2022 arXiv

  11. [19]

    3d local convolutional neural networks for gait recognition

    Zhen Huang, Dixiu Xue, Xu Shen, Xinmei Tian, Houqiang Li, Jianqiang Huang, and Xian-Sheng Hua. 3d local convolutional neural networks for gait recognition. In 2021 IEEE/CVF International Conference on Computer Vision (ICCV), pages 14900--14909, 2021

    2021

  12. [20]

    On-line fingerprint verification

    Anil Jain, Lin Hong, and Ruud Bolle. On-line fingerprint verification. IEEE transactions on pattern analysis and machine intelligence, 19 0 (4): 0 302--314, 1997

    1997

  13. [21]

    End-to-end semi-supervised learning for video action detection

    Akash Kumar and Yogesh Singh Rawat. End-to-end semi-supervised learning for video action detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

    2022

  14. [22]

    Benchmarking self-supervised video representation learning

    Akash Kumar, Ashlesha Kumar, Vibhav Vineet, and Yogesh Singh Rawat. Benchmarking self-supervised video representation learning. Neural Information Processing Systems 4th Workshop on Self-Supervised Learning: Theory and Practice, 2023

    2023

  15. [23]

    Contextual self-paced learning for weakly supervised spatio-temporal video grounding

    Akash Kumar, Zsolt Kira, and Yogesh Singh Rawat. Contextual self-paced learning for weakly supervised spatio-temporal video grounding. Proceedings of the International Conference on Learning Representations (ICLR), 2025 a

    2025

  16. [24]

    A large-scale analysis on contextual self-supervised video representation learning

    Akash Kumar, Ashlesha Kumar, Vibhav Vineet, and Yogesh S Rawat. A large-scale analysis on contextual self-supervised video representation learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 670--681, 2025 b

    2025

  17. [25]

    Stable mean teacher for semi-supervised video action detection

    Akash Kumar, Sirshapan Mitra, and Yogesh Singh Rawat. Stable mean teacher for semi-supervised video action detection. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 4419--4427, 2025 c

    2025

  18. [26]

    Facial attribute editing by latent space adversarial variational autoencoders

    Defang Li, Min Zhang, Weifu Chen, and Guocan Feng. Facial attribute editing by latent space adversarial variational autoencoders. In 2018 24th International Conference on Pattern Recognition (ICPR), pages 1337--1342. IEEE, 2018

    2018

  19. [27]

    Id3: Identity-preserving-yet-diversified diffusion models for synthetic face recognition

    Shen Li, Jianqing Xu, Jiaying Wu, Miao Xiong, Ailin Deng, Jiazhen Ji, Yuge Huang, Wenjie Feng, Shouhong Ding, and Bryan Hooi. Id3: Identity-preserving-yet-diversified diffusion models for synthetic face recognition. arXiv preprint arXiv:2409.17576, 2024

    2024 arXiv

  20. [28]

    Differ: Disentangling identity features via semantic cues for clothes-changing person re-id

    Xin Liang and Yogesh S Rawat. Differ: Disentangling identity features via semantic cues for clothes-changing person re-id. In Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR), pages 13980--13989, 2025

    2025

  21. [29]

    Gait recognition via effective global-local feature representation and local temporal aggregation

    Beibei Lin, Shunli Zhang, and Xin Yu. Gait recognition via effective global-local feature representation and local temporal aggregation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 14648--14656, 2021

    2021

  22. [30]

    Pedestrian attribute editing for gait recognition and anonymization

    Jingzhe Ma, Dingqiang Ye, Chao Fan, and Shiqi Yu. Pedestrian attribute editing for gait recognition and anonymization. arXiv preprint arXiv:2303.05076, 2023 a

    2023 arXiv

  23. [31]

    Dynamic aggregated network for gait recognition

    Kang Ma, Ying Fu, Dezhi Zheng, Chunshui Cao, Xuecai Hu, and Yongzhen Huang. Dynamic aggregated network for gait recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 22076--22085, 2023 b

    2023

  24. [32]

    Video action detection: Analysing limitations and challenges

    Rajat Modi, Aayush Jung Rana, Akash Kumar, Praveen Tirupattur, Shruti Vyas, Yogesh Singh Rawat, and Mubarak Shah. Video action detection: Analysing limitations and challenges. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 4907--4916, 2022

    2022

  25. [33]

    Automatic recognition by gait

    Mark S Nixon and John N Carter. Automatic recognition by gait. Proceedings of the IEEE, 94 0 (11): 0 2013--2024, 2006

    2013

  26. [34]

    Colors see colors ignore: Clothes changing reid with color disentanglement

    Priyank Pathak and Yogesh S Rawat. Colors see colors ignore: Clothes changing reid with color disentanglement. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025 a

    2025

  27. [35]

    Coarse attribute prediction with task agnostic distillation for real world clothes changing reid

    Priyank Pathak and Yogesh S Rawat. Coarse attribute prediction with task agnostic distillation for real world clothes changing reid. In 36th British Machine Vision Conference 2025, BMVC 2025, Sheffield, UK, November 24-27, 2025 . BMVA Press, 2025 b

    2025

  28. [36]

    Hyperface: Generating synthetic face recognition datasets by exploring face embedding hypersphere

    Hatef Otroshi Shahreza and Sebastien Marcel. Hyperface: Generating synthetic face recognition datasets by exploring face embedding hypersphere. arXiv preprint arXiv:2411.08470, 2024

    2024 arXiv

  29. [37]

    Semi-supervised active learning for video action detection

    Ayush Singh, Aayush J Rana, Akash Kumar, Shruti Vyas, and Yogesh Singh Rawat. Semi-supervised active learning for video action detection. Proceedings of the AAAI Conference on Artificial Intelligence, 38 0 (5): 0 4891--4899, 2024

    2024

  30. [38]

    Diffused heads: Diffusion models beat gans on talking-face generation

    Michal Stypulkowski, Konstantinos Vougioukas, Sen He, Maciej Zieba, Stavros Petridis, and Maja Pantic. Diffused heads: Diffusion models beat gans on talking-face generation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 5091--5100, 2024

    2024

  31. [39]

    Multi-view large population gait dataset and its performance evaluation for cross-view gait recognition

    Noriko Takemura, Yasushi Makihara, Daigo Muramatsu, Tomio Echigo, and Yasushi Yagi. Multi-view large population gait dataset and its performance evaluation for cross-view gait recognition. IPSJ transactions on Computer Vision and Applications, 10: 0 1--14, 2018

    2018

  32. [40]

    Facial image inpainting with variational autoencoder

    Ching-Ting Tu and Yi-Fu Chen. Facial image inpainting with variational autoencoder. In 2019 2nd international conference of intelligent robotic and control engineering (IRCE), pages 119--122. IEEE, 2019

    2019

  33. [41]

    Hierarchical spatio-temporal representation learning for gait recognition

    Lei Wang, Bo Liu, Fangfang Liang, and Bin Wang. Hierarchical spatio-temporal representation learning for gait recognition. ArXiv, abs/2307.09856, 2023 a

    2023 arXiv

  34. [42]

    Dygait: Exploiting dynamic representations for high-performance gait recognition

    Ming-Zhen Wang, Xianda Guo, Beibei Lin, Tian Yang, Zhenguo Zhu, Lincheng Li, Shunli Zhang, and Xin Yu. Dygait: Exploiting dynamic representations for high-performance gait recognition. ArXiv, abs/2303.14953, 2023 b

    2023 arXiv

  35. [43]

    Styleganex: Stylegan-based manipulation beyond cropped aligned faces

    Shuai Yang, Liming Jiang, Ziwei Liu, and Chen Change Loy. Styleganex: Stylegan-based manipulation beyond cropped aligned faces. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 21000--21010, 2023

    2023

  36. [44]

    Towards pose robust face recognition

    Dong Yi, Zhen Lei, and Stan Z Li. Towards pose robust face recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3539--3545, 2013

    2013

  37. [45]

    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. In 18th international conference on pattern recognition (ICPR'06), pages 441--444. IEEE, 2006

    2006

  38. [46]

    Gaitgan: Invariant gait feature extraction using generative adversarial networks

    Shiqi Yu, Haifeng Chen, Edel B Garcia Reyes, and Norman Poh. Gaitgan: Invariant gait feature extraction using generative adversarial networks. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops, pages 30--37, 2017

    2017

  39. [47]

    García Reyes, Yongzhen Huang, and Norman Poh

    Shiqi Yu, Rijun Liao, Weizhi An, Haifeng Chen, Edel B. García Reyes, Yongzhen Huang, and Norman Poh. Gaitganv2: Invariant gait feature extraction using generative adversarial networks. Pattern Recognit., 87: 0 179--189, 2019

    2019

  40. [48]

    Show-1: Marrying pixel and latent diffusion models for text-to-video generation

    David Junhao Zhang, Jay Zhangjie Wu, Jia-Wei Liu, Rui Zhao, Lingmin Ran, Yuchao Gu, Difei Gao, and Mike Zheng Shou. Show-1: Marrying pixel and latent diffusion models for text-to-video generation. arXiv preprint arXiv:2309.15818, 2023

    2023 arXiv

  41. [49]

    Magicvideo: Efficient video generation with latent diffusion models

    Daquan Zhou, Weimin Wang, Hanshu Yan, Weiwei Lv, Yizhe Zhu, and Jiashi Feng. Magicvideo: Efficient video generation with latent diffusion models. arXiv preprint arXiv:2211.11018, 2022

    2022 arXiv

  42. [50]

    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. In Proceedings of the IEEE/CVF international conference on computer vision, pages 14789--14799, 2021

    2021

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.