Pith. sign in

REVIEW 4 major objections 5 minor 50 references

MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation

T0 review · 4 major / 5 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 MV-SSM claims that treating body joints as a spatial sequence—scanned bidirectionally by a state-space model on top of projective multi-view attention—lets a 3D pose estimator transfer to camera setups it never trained on, beating the trans

desk verdict MV-SSM is a genuinely new SSM-based multi-view 3D pose estimator with promising cross-camera generalization numbers, but the key ablation is confounded and there are no error bars or code; still worth refereeing. read the letter →

arxiv 2509.00649 v1 pith:SMO4WGQP submitted 2025-08-31 cs.CV cs.RO

Aviral Chharia , Wenbo Gou , Haoye Dong This is my paper · ORCID
classification cs.CVcs.RO
keywords 3Dhumanposeestimationmulti-viewgeometrystatespacemodelsMambaprojectiveattentioncross-datasetgeneralizationcamera-arrangement
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

MV-SSM is an attempt to show that state space models—sequence models originally built for text and audio—can serve as the core of multi-view 3D human pose estimation. The paper's idea is to treat the body's joints as an ordered sequence, not over time but over the joint dimension, and to scan the multi-view image features around each projected keypoint bidirectionally. Why this matters: if the reported numbers hold, a pose estimator trained on one five-camera rig keeps working—and often improves over attention-based transformers—when cameras are removed, rearranged, or replaced by different datasets, a failure mode of earlier methods. The evidence is a set of in-domain and generalization results on standard benchmarks, plus ablations showing that removing the state-space component costs 5.8 points of AP25.

What carries the argument

The PSS block is the architectural unit that carries the paper's argument. Its projective-attention branch projects the current 3D keypoint estimate into each view, samples deformable points around the projected anchor, and aggregates local image features; its state-space branch takes the J joint tokens of each person and scans them as a sequence, not along time but along the joint dimension. GTBS is the scan variant: instead of scanning all image patches, it scans the sampled projection tokens bidirectionally, so the SSM sees relevant local tokens from every view while ignoring background. Stacked PSS blocks refine the keypoint tokens and output 2D residual offsets; a differentiable triangu

What would settle it

Retrain MV-SSM twice with identical settings except that the joint scan order in GTBS is the fixed skeleton-index order in one run and a fixed random permutation in the other; if AP25 on the CMU Panoptic test split does not fall in the random-order run, the sequential state-space modeling is not what produces the reported gains and the contribution reduces to any-order token mixing.

Watch

Extended reading notes

Core claim

The paper's central claim is that a multi-view 3D pose estimator can be built by modeling the ordered list of body joints as a spatial sequence at two levels: the visual-feature level (sampled multi-view image tokens) and the person-keypoint level (explicit 3D joint positions). The PSS block does this by coupling projective attention—which gathers image features around each joint's projected location in each view—with a Mamba-style state-space layer that scans the joint tokens bidirectionally (GTBS). The paper reports that this combination reaches 93.5 AP25 and 15.7 mm MPJPE on CMU Panoptic, above MVGFormer's 92.3/16.0, and that the margin grows when cameras are removed, rearranged, or the m

Load-bearing premise

The load-bearing premise is that the joint tokens form a meaningful spatial sequence with a definite order; the paper never specifies that order, so the state-space layer's sequential dependencies could be artifacts of the skeleton's fixed index rather than learned geometric structure.

Editorial extensions

If this is right

  • If the reported results hold, multi-view 3D pose estimation no longer needs to be framed exclusively as an attention problem: a linear-time state-space scan over joint tokens can carry the inter-joint modeling and matches or beats the transformer baseline in-domain.
  • A model trained on one five-camera rig transfers to three, four, six, and seven cameras and to different camera IDs without fine-tuning, so practitioners could train once and deploy on differently sized capture volumes.
  • The ablation numbers make the SSM component load-bearing: dropping GTBS and Mamba together lowers AP25 from 93.5 to 87.7 and raises MPJPE from 15.7 mm to 18.6 mm.
  • The explicit 3D keypoint branch matters more than the visual-token branch for accuracy (91.8 vs 92.5 AP25 when each is removed), pointing to geometry-anchored refinement as the main driver.
  • The biggest gain over MVGFormer appears with only three cameras (+10.8 AP25), suggesting the sequence model is most valuable when each view is information-poor.

Reading between the lines

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

  • A testable next step the paper leaves open is to define the joint scan order kinematically (e.g., along the skeleton chain or by limb hierarchy) and compare it with the dataset's index order; this would separate true geometric sequencing from token-mixing artifacts.
  • Because GTBS operates only on projected anchor tokens, the same PSS block could be reused as a refinement head in other differentiable multi-view geometric pipelines, including object keypoint or camera-pose estimation, without retraining the whole backbone.
  • If the cross-arrangement gains come mainly from projective attention's local sampling, then varying the sampling radius and token count while freezing the SSM would reveal how much of the generalization is geometric versus sequential.
  • The paper's own failure case—a person visible in only one camera—suggests a straightforward extension: feed the PSS block with synthesized or hallucinated missing-view tokens so the state-space layer can propagate evidence across joints even when one view is absent.
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

4 major / 5 minor

Summary. The paper proposes MV-SSM, an end-to-end multi-view 3D human pose estimation framework built around a Projective State Space (PSS) block that combines projective attention with Mamba-style state space modeling, and a Grid Token-guided Bidirectional Scanning (GTBS) scheme. The model is trained on CMU Panoptic and evaluated in-domain and on cross-camera and cross-scene generalization benchmarks (CMU Panoptic variants, Campus, Shelf). The central claim is that MV-SSM generalizes to unseen camera configurations better than the transformer-based MVGFormer, with reported gains of +10.8 AP25 on the three-camera setting, +7.0 AP25 on a challenging arrangement, and +15.3 PCP on Campus Actor 1, while matching in-domain accuracy.

Significance. If the claims hold, this would be a useful contribution: it is an early adaptation of state space models to multi-view 3D pose estimation, and the cross-camera generalization results are practically relevant. The paper includes external benchmark comparisons, a clear architectural description, and ablation experiments on the proposed components. The main weaknesses are that the central mechanism is not cleanly isolated by the ablations, the reported margins are not accompanied by variance estimates, and a key design choice (the joint spatial sequence ordering) is left unspecified. These issues are load-bearing for the paper's strongest claims, but they are addressable with additional experiments and clarification.

major comments (4)
  1. [§4.2, Table 4 Row 4; Supplementary Section 3] The ablation labeled 'w/o GTBS + Mamba' is confounded. The supplementary states that this variant 'modified the appearance token to encode only the instance-level information' and that 'multi-view feature fusion also degenerates to only fuse the instance level features.' Thus Row 4 changes the token granularity and fusion level while also removing GTBS and Mamba. The 93.5→87.7 AP25 drop cannot be attributed to the proposed scanning mechanism. Since Row 3, which removes only Mamba, shows a 1.2 AP25 drop, the paper's statement that GTBS and Mamba are 'integral' is not established by this table. Moreover, all ablations are on the in-domain CMU Panoptic split; no component ablation isolates the contribution of PSS/GTBS to the cross-camera and cross-scene generalization gains in Tables 2 and 3.
  2. [Tables 1–4] No repeated runs, error bars, or significance tests are reported. Several comparisons to MVGFormer are small: Table 1 reports 93.5 vs. 92.3 AP25; Table 3 reports CMU1 86.0 vs. 86.8 AP25 (MV-SSM is lower on AP25) and CMU0(6) 94.6 vs. 94.7 AP25. Without variance estimates, these margins may be within run-to-run noise. Please report at least three seeds with mean and standard deviation for the central comparisons.
  3. [§3.1, Figure 2] The 'joint spatial sequence' is not defined. The text says the PSS block 'scans along joint dimension rather than the original time dimension,' but does not specify the ordering of the J joints in the sequence. If the default order is simply the skeleton index order, the sequential dependencies learned by the SSM may be indexing artifacts. Please specify how the joint sequence is constructed (e.g., a fixed kinematic order, a learned order, or a spatial sorting) and ablate at least one alternative ordering to show the mechanism is not an artifact of the chosen sequence.
  4. [Supplementary Section 2.2; Table 2] For cross-dataset evaluations on Campus and Shelf, the paper reports different 3D search-space centers for each dataset ([2000,5000,1000] for Campus, [0,500,800] for Shelf), while claiming the model is evaluated 'w/o fine-tuning.' If these centers are chosen using dataset-specific knowledge, the cross-dataset generalization claim is partially weakened. Please clarify whether these values are fixed a priori, chosen by a common automatic rule, or manually tuned, and discuss sensitivity of the cross-dataset results to this hyperparameter.
minor comments (5)
  1. [Eqs. (1)–(2)] The notation switches between continuous-time dynamics (h'(t)) and discrete-time recurrence (ht) without explicit discretization steps; also y(t) appears as Ch(t)+Dx(t) in Eq. (1) and Ch(t) in Eq. (2). Please align notation.
  2. [Table 3] The heading 'Cross-arrange' should read 'Cross-arrangement' for consistency.
  3. [Table 4] Row 4 label 'w/o GTBS + Mamba' is ambiguous; the supplementary clarifies that the ablation also changes token granularity. The main text should state this explicitly.
  4. [Supplementary Eq. (7)] The AP formula uses k = 0 to n−1 with r(k) − r(k+1), which would require an r(n) term. Please clarify the indexing or define n.
  5. [Introduction] Typo: 'off-the-self models' should be 'off-the-shelf models.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; MV-SSM is an empirical architecture paper whose main claims are validated on external benchmarks.

full rationale

MV-SSM does not derive its results from a fitted parameter or a self-citation chain. The central accuracy and generalization claims (abstract; Tables 1-3) are evaluated on CMU Panoptic, Campus, and Shelf, with no fine-tuning on the target datasets, so the reported 'predictions' are not forced by construction. The PSS block and GTBS are architectural proposals; Section 3.1's statement that the block 'scans along joint dimension rather than the original time dimension' is under-specified regarding the joint ordering, but that is a reproducibility/mechanism concern, not a circular reduction. Table 4 Row 4 ('w/o GTBS + Mamba') is a confounded ablation: Supplementary Section 3 says it 'modified the appearance token to encode only the instance-level information' and 'multi-view feature fusion also degenerates to only fuse the instance level features,' so the 5.8 AP25 drop cannot be uniquely attributed to removing GTBS+Mamba. This weakens the component-attribution claim, but it is an experimental-control issue, not an instance of a prediction reducing to its input by definition; no fitted parameter is renamed as a prediction and no result is imported from the authors' own prior work. The Conclusion's limitation ('MV-SSM fails when a human is visible in only one camera') is an acknowledged boundary condition, not a circular step. No self-citations by the present authors are load-bearing; references to Mamba, VMamba, MvP, and MVGFormer are external prior art. Consequently, the circularity score is 0.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The paper is entirely empirical: there are no derivations, so the ledger items are hand-chosen constants and background assumptions rather than fitted derivation inputs. The dataset volumes, initial token count, and confidence threshold are field-standard hyperparameters but qualify as free parameters because the generalization claims depend on them and they were chosen with knowledge of all evaluation benchmarks. No new physical entities are introduced; the only new construct, GTBS, has ablation support that is partial and conflated.

free parameters (5)
  • Per-dataset space size and center = CMU Panoptic [8000,8000,2000] mm at [0,-500,800]; Campus [2000,5000,1000]; Shelf [0,500,800]
    Hand-chosen coordinate volumes (supplementary Section 2.2) define where the initial T-pose tokens are sampled and where predictions are filtered; the cross-dataset evaluations reuse the CMU-trained model with these per-dataset constants, so they encode prior knowledge of all three test benchmarks.
  • Initial token count N = ~1024
    Hand-chosen 'based on the motion capture space of the dataset' (supplementary Section 1); performance depends on the token budget covering all persons in the volume.
  • Confidence threshold epsilon = 0.1
    Layer-wise filtering threshold for the positive/negative token classifier (Section 4, Implementation Details); affects which tokens survive to refinement and triangulation.
  • Number of PSS/decoder layers M = 4
    Stack depth of the progressive refinement (supplementary Section 1); never ablated.
  • W nearest-token assignment count = not reported
    Eq. 6 assigns the W nearest tokens to each ground-truth person in the loss; W is never stated, and the training loss depends on it.
assumptions (5)
  • domain assumption Camera poses (projection matrices) are known and fixed during training and inference.
    Section 4, Evaluation Metrics: 'The camera pose is assumed to be known and utilized, following common practice.' All triangulation and projective attention depend on this.
  • domain assumption The pinhole projection model and differentiable algebraic triangulation of Iskakov et al. [20] yield valid 3D estimates from 2D predictions and confidences.
    Used in Eq. 4 and throughout; inherited machinery, not re-derived.
  • standard math Zero-order hold discretization (Eq. 3) correctly bridges the continuous SSM (Eq. 1) to the discrete recurrence (Eq. 2).
    Section 3.1 Preliminaries; standard SSM discretization from the S4/Mamba literature [15,16].
  • domain assumption A COCO-pretrained ResNet-50 backbone provides multi-scale features that transfer to multi-view pose estimation.
    Section 4, Implementation Details; backbones are frozen with the same weights as in MvP, VoxelPose, MVGFormer.
  • domain assumption Anchor-based matching [46] reliably assigns ground-truth poses to the nearest initial tokens, and the ground-plane T-pose initializations sufficiently cover all subjects in a scene.
    Section 3.2 loss formulation; the learned token classifier and progressive refinement depend on this assignment being correct.
invented entities (1)
  • Grid Token-guided Bidirectional Scanning (GTBS)
    purpose: Bidirectional SSM scanning over sampled projection tokens at both the visual-feature and person-keypoint levels, replacing naive full-image scanning inside the PSS block.
    The only evidence is the within-paper ablation (Table 4, Row 4: AP25 drops 93.5 to 87.7, MPJPE rises 15.7 to 18.6 mm), and that ablation is conflated because GTBS is removed together with the Mamba blocks and a token-encoding change. No independent external prediction is offered for GTBS in isolation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation." pith.science (2026). https://pith.science/paper/SMO4WGQP

@misc{pith2026250900649,
  author       = {Pith},
  title        = {Pith review of: MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SMO4WGQP}},
  note         = {Machine review of arXiv:2509.00649}
}
read the original abstract

While significant progress has been made in single-view 3D human pose estimation, multi-view 3D human pose estimation remains challenging, particularly in terms of generalizing to new camera configurations. Existing attention-based transformers often struggle to accurately model the spatial arrangement of keypoints, especially in occluded scenarios. Additionally, they tend to overfit specific camera arrangements and visual scenes from training data, resulting in substantial performance drops in new settings. In this study, we introduce a novel Multi-View State Space Modeling framework, named MV-SSM, for robustly estimating 3D human keypoints. We explicitly model the joint spatial sequence at two distinct levels: the feature level from multi-view images and the person keypoint level. We propose a Projective State Space (PSS) block to learn a generalized representation of joint spatial arrangements using state space modeling. Moreover, we modify Mamba's traditional scanning into an effective Grid Token-guided Bidirectional Scanning (GTBS), which is integral to the PSS block. Multiple experiments demonstrate that MV-SSM achieves strong generalization, outperforming state-of-the-art methods: +10.8 on AP25 (+24%) on the challenging three-camera setting in CMU Panoptic, +7.0 on AP25 (+13%) on varying camera arrangements, and +15.3 PCP (+38%) on Campus A1 in cross-dataset evaluations. Project Website: https://aviralchharia.github.io/MV-SSM

Figures

Figures reproduced from arXiv: 2509.00649 by the authors.

Figure 1
Figure 1. Motivation. Comparison of different token scanning methods. (a) Cross Attention acts on all image tokens. (b) Projec￾tive Attention [27, 44] obtains anchors with perspective projection and selectively attends to sample tokens surrounding the anchor points. (c) The proposed Grid Token-guided Bidirectional Scan￾ning (GTBS) encodes the local context and the joint spatial se￾quence at the visual feature and person-keypo… view at source ↗
Figure 2
Figure 2. The architecture of the proposed Multi-view State Space Model ( [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The architecture of Mamba block [15], VSS block [29] and the proposed Projective State Space block. The PSS block captures joint spatial relationships through projective attention and state-space modeling, progressively refining results. lated as follows: k ′ = AlgTriangulation(u ′ t, ct, Πt), (4) where u ′ t is the 2D positions in T views, ct is correspond￾ing confidence scores, and Πt is the projection matrix for … view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Qualitative Comparisons. We present a visual compar￾ison against MVGFormer [27] on CMU Panoptic [21] benchmark. The Ground truth human poses are shown in “red” and the pre￾dicted pose is overlapped on it to show an accurate comparison. MV-SSM achieves accurate poses, e…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 46 canonical work pages

  1. [1]

    Pose-conditioned joint an- gle limits for 3d human pose reconstruction

    Ijaz Akhter and Michael J Black. Pose-conditioned joint an- gle limits for 3d human pose reconstruction. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 1446–1455, 2015. 1, 2

    work page 2015

  2. [2]

    Generalizable human pose triangulation

    Kristijan Bartol, David Bojani ´c, Tomislav Petkovi ´c, and Tomislav Pribani´c. Generalizable human pose triangulation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11028–11037, 2022. 6

    work page 2022

  3. [3]

    3d pictorial structures for multiple human pose estimation

    Vasileios Belagiannis, Sikandar Amin, Mykhaylo Andriluka, Bernt Schiele, Nassir Navab, and Slobodan Ilic. 3d pictorial structures for multiple human pose estimation. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 1669–1676, 2014. 6, 7, 11

    work page 2014

  4. [4]

    Multi-person 3d pose estimation and tracking in sports

    Lewis Bridgeman, Marco V olino, Jean-Yves Guillemaut, and Adrian Hilton. Multi-person 3d pose estimation and tracking in sports. In Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition workshops, pages 0–0, 2019. 1

    work page 2019

  5. [5]

    Openpose: Realtime multi-person 2d pose estimation using part affinity fields, 2019

    Zhe Cao, Gines Hidalgo, Tomas Simon, Shih-En Wei, and Yaser Sheikh. Openpose: Realtime multi-person 2d pose estimation using part affinity fields, 2019. 1

    work page 2019

  6. [6]

    Multi-person 3d pose estimation in crowded scenes based on multi-view geometry

    He Chen, Pengfei Guo, Pengfei Li, Gim Hee Lee, and Gre- gory Chirikjian. Multi-person 3d pose estimation in crowded scenes based on multi-view geometry. In Computer Vision– ECCV 2020: 16th European Conference, Glasgow, UK, Au- gust 23–28, 2020, Proceedings, Part III 16, pages 541–557. Springer, 2020. 1

    work page 2020

  7. [7]

    Cross-view tracking for multi-human 3d pose estimation at over 100 fps

    Long Chen, Haizhou Ai, Rui Chen, Zijie Zhuang, and Shuang Liu. Cross-view tracking for multi-human 3d pose estimation at over 100 fps. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 3279–3288, 2020

    work page 2020

  8. [8]

    Part-aware measurement for robust multi-view multi-human 3d pose estimation and track- ing

    Hau Chu, Jia-Hong Lee, Yao-Chih Lee, Ching-Hsien Hsu, Jia-Da Li, and Chu-Song Chen. Part-aware measurement for robust multi-view multi-human 3d pose estimation and track- ing. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 1472–1481,

    work page

Show all 50 references
  1. [9]

    Learning 3d human pose from structure and motion

    Rishabh Dabral, Anurag Mundhada, Uday Kusupati, Safeer Afaque, Abhishek Sharma, and Arjun Jain. Learning 3d human pose from structure and motion. In Proceedings of the European conference on computer vision (ECCV), pages 668–683, 2018. 1, 2

    2018

  2. [10]

    Convolutional neural networks on graphs with fast localized spectral filtering

    Micha ¨el Defferrard, Xavier Bresson, and Pierre Van- dergheynst. Convolutional neural networks on graphs with fast localized spectral filtering. Advances in neural informa- tion processing systems, 29, 2016. 2

    2016

  3. [11]

    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. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 3

    2009

  4. [12]

    Fast and robust multi-person 3d pose estima- tion from multiple views

    Junting Dong, Wen Jiang, Qixing Huang, Hujun Bao, and Xiaowei Zhou. Fast and robust multi-person 3d pose estima- tion from multiple views. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 7792–7801, 2019. 1, 2, 6, 7

    2019

  5. [13]

    Fast and robust multi-person 3d pose estimation and tracking from multiple views

    Junting Dong, Qi Fang, Wen Jiang, Yurou Yang, Qixing Huang, Hujun Bao, and Xiaowei Zhou. Fast and robust multi-person 3d pose estimation and tracking from multiple views. IEEE transactions on pattern analysis and machine intelligence, 44(10):6981–6992, 2021. 1

    2021

  6. [14]

    Pose locality constrained representation for 3d human pose reconstruction

    Xiaochuan Fan, Kang Zheng, Youjie Zhou, and Song Wang. Pose locality constrained representation for 3d human pose reconstruction. In Computer Vision–ECCV 2014: 13th Eu- ropean Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part I 13 , pages 174–188. Springer,

    2014

  7. [15]

    Mamba: Linear-time sequence mod- eling with selective state spaces

    Albert Gu and Tri Dao. Mamba: Linear-time sequence mod- eling with selective state spaces. 2023. 2, 3, 4, 5

    2023

  8. [16]

    Efficiently mod- eling long sequences with structured state spaces

    Albert Gu, Karan Goel, and Christopher R´e. Efficiently mod- eling long sequences with structured state spaces. 2021. 2, 3

    2021

  9. [17]

    Combining recurrent, convolutional, and continuous-time models with linear state space layers

    Albert Gu, Isys Johnson, Karan Goel, Khaled Saab, Tri Dao, Atri Rudra, and Christopher R ´e. Combining recurrent, convolutional, and continuous-time models with linear state space layers. Advances in neural information processing sys- tems, 34:572–585, 2021. 2

    2021

  10. [18]

    Mambavision: A hybrid mamba-transformer vision backbone, 2024

    Ali Hatamizadeh and Jan Kautz. Mambavision: A hybrid mamba-transformer vision backbone, 2024. 3

    2024

  11. [19]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 2, 3

    2016

  12. [20]

    Learnable triangulation of human pose

    Karim Iskakov, Egor Burkov, Victor Lempitsky, and Yury Malkov. Learnable triangulation of human pose. InProceed- ings of the IEEE/CVF international conference on computer vision, pages 7718–7727, 2019. 1, 2, 3, 4

    2019

  13. [21]

    Panoptic studio: A massively multiview system for social motion capture

    Hanbyul Joo, Hao Liu, Lei Tan, Lin Gui, Bart Nabbe, Iain Matthews, Takeo Kanade, Shohei Nobuhara, and Yaser Sheikh. Panoptic studio: A massively multiview system for social motion capture. In Proceedings of the IEEE inter- national conference on computer vision , pages 3334–3342,

  14. [22]

    A new approach to linear filtering and prediction problems

    Rudolph Emil Kalman. A new approach to linear filtering and prediction problems. 1960. 2

    1960

  15. [23]

    Videomamba: State space model for efficient video understanding

    Kunchang Li, Xinhao Li, Yi Wang, Yinan He, Yali Wang, Limin Wang, and Yu Qiao. Videomamba: State space model for efficient video understanding. In European Conference on Computer Vision, pages 237–255. Springer, 2025. 3

    2025

  16. [24]

    3d human pose estimation from monocular images with deep convolutional neural network

    Sijin Li and Antoni B Chan. 3d human pose estimation from monocular images with deep convolutional neural network. In Computer Vision–ACCV 2014: 12th Asian Conference on Computer Vision, Singapore, Singapore, November 1-5, 2014, Revised Selected Papers, Part II 12 , pages 332–...

    2014

  17. [25]

    Mamba- nd: Selective state space modeling for multi-dimensional data

    Shufan Li, Harkanwar Singh, and Aditya Grover. Mamba- nd: Selective state space modeling for multi-dimensional data. In European Conference on Computer Vision , pages 75–92. Springer, 2025. 3

    2025

  18. [26]

    Mhformer: Multi-hypothesis transformer for 3d 9 human pose estimation

    Wenhao Li, Hong Liu, Hao Tang, Pichao Wang, and Luc Van Gool. Mhformer: Multi-hypothesis transformer for 3d 9 human pose estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 13147–13156, 2022. 1, 2

    2022

  19. [27]

    Multiple view geometry transform- ers for 3d human pose estimation

    Ziwei Liao, Jialiang Zhu, Chunyu Wang, Han Hu, and Steven L Waslander. Multiple view geometry transform- ers for 3d human pose estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 708–717, 2024. 1, 2, 3, 4, 5, 6, 7, 8, 11, 12

    2024

  20. [28]

    Multi-view multi-person 3d pose estimation with plane sweep stereo

    Jiahao Lin and Gim Hee Lee. Multi-view multi-person 3d pose estimation with plane sweep stereo. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11886–11895, 2021. 6

    2021

  21. [29]

    Vmamba: Visual state space model

    Yue Liu, Yunjie Tian, Yuzhong Zhao, Hongtian Yu, Lingxi Xie, Yaowei Wang, Qixiang Ye, and Yunfan Liu. Vmamba: Visual state space model. 2024. 2, 3, 4, 5

    2024

  22. [30]

    Context modeling in 3d human pose estimation: A unified perspective

    Xiaoxuan Ma, Jiajun Su, Chunyu Wang, Hai Ci, and Yizhou Wang. Context modeling in 3d human pose estimation: A unified perspective. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 6238–6247, 2021. 2

    2021

  23. [31]

    A simple yet effective baseline for 3d human pose esti- mation

    Julieta Martinez, Rayat Hossain, Javier Romero, and James J Little. A simple yet effective baseline for 3d human pose esti- mation. In Proceedings of the IEEE international conference on computer vision, pages 2640–2649, 2017. 1, 2

    2017

  24. [32]

    Coarse-to-fine volumetric pre- diction for single-image 3d human pose

    Georgios Pavlakos, Xiaowei Zhou, Konstantinos G Derpa- nis, and Kostas Daniilidis. Coarse-to-fine volumetric pre- diction for single-image 3d human pose. In Proceedings of the IEEE conference on computer vision and pattern recog- nition, pages 7025–7034, 2017. 2

    2017

  25. [33]

    Matching and re- covering 3d people from multiple views

    Alejandro Perez-Yus and Antonio Agudo. Matching and re- covering 3d people from multiple views. In Proceedings of the IEEE/CVF Winter Conference on Applications of Com- puter Vision, pages 3622–3631, 2022. 1

    2022

  26. [34]

    Re- constructing 3d human pose from 2d image landmarks

    Varun Ramakrishna, Takeo Kanade, and Yaser Sheikh. Re- constructing 3d human pose from 2d image landmarks. In Computer Vision–ECCV 2012: 12th European Conference on Computer Vision, Florence, Italy, October 7-13, 2012, Proceedings, Part IV 12, pages 573–586. Springer, 2012. 1, 2

    2012

  27. [35]

    Tessetrack: End-to- end learnable multi-person articulated 3d pose tracking

    N Dinesh Reddy, Laurent Guigues, Leonid Pishchulin, Jayan Eledath, and Srinivasa G Narasimhan. Tessetrack: End-to- end learnable multi-person articulated 3d pose tracking. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 15190–15200...

    2021

  28. [36]

    V oxelpose: Towards multi-camera 3d human pose estimation in wild en- vironment

    Hanyue Tu, Chunyu Wang, and Wenjun Zeng. V oxelpose: Towards multi-camera 3d human pose estimation in wild en- vironment. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceed- ings, Part I 16 , pages 197–212. Springer, 2020. 1, 2, 6, 7, 11

    2020

  29. [37]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017. 2

    2017

  30. [38]

    Yolov10: Real-time end- to-end object detection

    Ao Wang, Hui Chen, Lihao Liu, Kai Chen, Zijia Lin, Jun- gong Han, and Guiguang Ding. Yolov10: Real-time end- to-end object detection. arXiv preprint arXiv:2405.14458 ,

    arXiv

  31. [39]

    Robust estimation of 3d human poses from a single image

    Chunyu Wang, Yizhou Wang, Zhouchen Lin, Alan L Yuille, and Wen Gao. Robust estimation of 3d human poses from a single image. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2361–2368,

  32. [40]

    Graph-based 3d multi-person pose estimation using multi-view images

    Size Wu, Sheng Jin, Wentao Liu, Lei Bai, Chen Qian, Dong Liu, and Wanli Ouyang. Graph-based 3d multi-person pose estimation using multi-view images. In Proceedings of the IEEE/CVF international conference on computer vision, pages 11148–11157, 2021. 6

    2021

  33. [41]

    Simple baselines for human pose estimation and tracking

    Bin Xiao, Haiping Wu, and Yichen Wei. Simple baselines for human pose estimation and tracking. In Proceedings of the European conference on computer vision (ECCV), pages 466–481, 2018. 11

    2018

  34. [42]

    A dual-source approach for 3d pose es- timation from a single image

    Hashim Yasin, Umar Iqbal, Bjorn Kruger, Andreas Weber, and Juergen Gall. A dual-source approach for 3d pose es- timation from a single image. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 4948–4956, 2016. 1, 2

    2016

  35. [43]

    Faster voxelpose: Real-time 3d human pose estima- tion by orthographic projection

    Hang Ye, Wentao Zhu, Chunyu Wang, Rujie Wu, and Yizhou Wang. Faster voxelpose: Real-time 3d human pose estima- tion by orthographic projection. In European Conference on Computer Vision, pages 142–159. Springer, 2022. 1, 6

    2022

  36. [44]

    Direct multi-view multi-person 3d pose estimation

    Jianfeng Zhang, Yujun Cai, Shuicheng Yan, Jiashi Feng, et al. Direct multi-view multi-person 3d pose estimation. Advances in Neural Information Processing Systems , 34: 13153–13164, 2021. 1, 2, 4, 5, 6, 7, 11, 12

    2021

  37. [45]

    4d association graph for realtime multi-person motion capture using multiple video cameras

    Yuxiang Zhang, Liang An, Tao Yu, Xiu Li, Kun Li, and Yebin Liu. 4d association graph for realtime multi-person motion capture using multiple video cameras. In Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1324–1333, 2020. 1

    2020

  38. [46]

    Ob- jects as points

    Xingyi Zhou, Dequan Wang, and Philipp Kr ¨ahenb¨uhl. Ob- jects as points. arXiv preprint arXiv:1904.07850, 2019. 5

    1904 arXiv

  39. [47]

    Vision mamba: Efficient visual representation learning with bidirectional state space model

    Lianghui Zhu, Bencheng Liao, Qian Zhang, Xinlong Wang, Wenyu Liu, and Xinggang Wang. Vision mamba: Efficient visual representation learning with bidirectional state space model. arXiv preprint arXiv:2401.09417, 2024. 2, 3 10 MV-SSM: Multi-View State Space Modeling for 3D Human...

    2024 arXiv

  40. [48]

    We used the same backbone weights as previous works [27, 36, 44] for a fair comparison

    Model Architecture Details The training images, sized 960 × 512, from the CMU- Panoptic dataset were fed into a ResNet-50 backbone pre-trained on the COCO dataset for 2D pose estimation task [41]. We used the same backbone weights as previous works [27, 36, 44] for a fair comp...

  41. [49]

    Datasets We outline the details of the datasets employed in this sec- tion

    Training and other details 2.1. Datasets We outline the details of the datasets employed in this sec- tion. These include the CMU Panoptic [21], Shelf [3], and Campus [3] datasets. Note that only CMU Panoptic was used for training. • CMU Panoptic [21] is a 3D multi-view datase...

    2000

  42. [50]

    Since the results of the ablations have already been dis- cussed in detail in the main paper in Rows 1-6 of Table 4, we focus on an experiment-wise detailed description

    Ablation Details We provide an in-depth description of the modifications made to MV-SSM to perform the ablation experiments. Since the results of the ablations have already been dis- cussed in detail in the main paper in Rows 1-6 of Table 4, we focus on an experiment-wise deta...

Pith tools

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