Abstract
Parametric methods are widely utilized in RGB-based human mesh recovery, relying on precise statistical human body model parameters that are challenging to obtain. In contrast, non-parametric transformer-based approaches excel but are applied only to monocular RGB tasks. To address these limitations, this paper presents Semi-Supervised Multi-View Human Mesh Recovery Transformer (SS-MVMETRO), which combines multi-view information with non-parametric methods for the first time. Our model encodes different images according to their respective view directions, fusing local features around key points of joints and vertices. Then, a residual-like structure is proposed to integrate the fused features in the mesh recovery transformer, which subsequently predicts the 3D coordinates of the human mesh vertices. Additionally, we divide different views into the main view and auxiliary views and propose a semi-supervised training approach that requires fewer matching labels. The efficacy of our work is validated on two datasets, Human3.6M and Mpi_inf_3dph, through quantitative and qualitative experiments. The results indicate that SS-MVMETRO improves the reconstruction accuracy, surpassing previous image-based methods by at least 8.9% in terms of Procrustes Analysis Mean-Per-Joint-Position-Error (PA-MPJPE).
Graphical abstract
Similar content being viewed by others
Data availability and access
The “Human3.6M” data that support the findings of this work are available in Human3.6M, the ”Mpi_inf_3dph” data are available in Mpi_inf_3dph, the ”Mpii” data are available in Mpii, the ”Muco” data are available in Muco, the ”Up3d” data are available in Up3d, the ”Coco” data are available in Coco. These datasets are publicly accessible.
References
Loper M, Mahmood N, Romero J et al (2015) Smpl: A skinned multi-person linear model. ACM Transactions on Graphics 34(6):1–16. https://doi.org/10.1145/2816795.2818013
Ran H, Ning X, Li W et al (2023) 3d human pose and shape estimation via de-occlusion multi-task learning. Neurocomputing 126284. https://doi.org/10.1016/j.neucom.2023.126284
Wei G, Lan C, Zeng W et al (2020) View invariant 3d human pose estimation. IEEE Trans Circuits Syst Video Technol 30(12):4601–4610. https://doi.org/10.1109/TCSVT.2019.2928813
Gu R, Wang G, Jiang Z et al (2020) Multi-person hierarchical 3d pose estimation in natural videos. IEEE Trans Circuits Syst Video Technol 30(11):4245–4257. https://doi.org/10.1109/TCSVT.2019.2953678
Kolotouros N, Pavlakos G, Black MJ et al (2019) Learning to reconstruct 3d human pose and shape via model-fitting in the loop. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 2252–2261
Zhang H, Tian Y, Zhou X et al (2021) Pymaf: 3d human pose and shape regression with pyramidal mesh alignment feedback loop. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 11446–11456
Liang J, Lin MC (2019) Shape-aware human pose and shape reconstruction using multi-view images. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 4352–4362
Lin K, Wang L, Liu Z (2021) Mesh graphormer. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 12939–12948
Cho J, Youwang K, Oh TH (2022) Cross-attention of disentangled modalities for 3d human mesh recovery with transformers. In: Proceedings of the European conference on computer vision, Springer, pp 342–359
Dong Y, Yuan Q, Peng R et al (2024) An iterative 3d human body reconstruction method driven by personalized dimensional prior knowledge. Appl Intell 54(1):738–748. https://doi.org/10.1007/s10489-023-05214-y
Kim J, Gwon MG, Park H et al (2023) Sampling is matter: Point-guided 3d human mesh reconstruction. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 12880–12889
Dai Y, Wen C, Wu H et al (2022) Indoor 3d human trajectory reconstruction using surveillance camera videos and point clouds. IEEE Trans Circuits Syst Video Technol 32(4):2482–2495. https://doi.org/10.1109/TCSVT.2021.3081591
Zhang B, Ma K, Wu S et al (2023) Two-stage co-segmentation network based on discriminative representation for recovering human mesh from videos. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 5662–5670
Zheng Z, Yu T, Liu Y et al (2022) Pamir: Parametric model-conditioned implicit representation for image-based human reconstruction. IEEE Trans Pattern Anal Mach Intell 44(6):3170–3184. https://doi.org/10.1109/TPAMI.2021.3050505
Harvey FG, Yurick M, Nowrouzezahrai D et al (2020) Robust motion in-betweening. ACM Trans Graphics (TOG) 39(4):60–1. https://doi.org/10.1145/3386569.3392480
Henter GE, Alexanderson S, Beskow J (2020) Moglow: Probabilistic and controllable motion synthesis using normalising flows. ACM Trans Graphics (TOG) 39(6):1–14. https://doi.org/10.1145/3414685.3417836
Tian Y, Zhang H, Liu Y et al (2023) Recovering 3d human mesh from monocular images: A survey. IEEE Trans Pattern Anal Mach Intell 45(12):15406–15425. https://doi.org/10.1109/TPAMI.2023.3298850
Bogo F, Kanazawa A, Lassner C et al (2016) Keep it smpl: Automatic estimation of 3d human pose and shape from a single image. In: Proceedings of the European conference on computer vision, Springer, pp 561–578
Mahendran S, Ali H, Vidal R (2018) A mixed classification-regression framework for 3d pose estimation from 2d images. In: Proceedings of the British machine vision conference. BMVA Press, pp 72–84
Lin K, Wang L, Liu Z (2021) End-to-end human pose and mesh reconstruction with transformers. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 1954–1963
Shin S, Halilaj E (2020) Multi-view human pose and shape estimation using learnable volumetric aggregation
Li Z, Oskarsson M, Heyden A (2021) 3d human pose and shape estimation through collaborative learning and multi-view model-fitting. In: Proceedings of the IEEE/CVF winter conference on applications of computer vision, pp 1888–1897
Zhang S, Liu Y, Liu J et al (2022) Multi-view high precise 3d human body reconstruction method for virtual fitting. Int J Pattern Recognition Artif Intell 36(15):2256023. https://doi.org/10.1142/S0218001422560237
Vaswani A, Shazeer N, Parmar N et al (2017) Attention is all you need. In: Proceedings of the International Conference on neural information processing systems, pp 6000–6010
Zhang J, Cai Y, Yan S et al (2021) Direct multi-view multi-person 3d pose estimation. In: Proceedings of the International Conference on neural information processing systems, pp 13153–13164
Hao C, Kong D, Li J et al (2023) Hypergraph based human mesh hierarchical representation and reconstruction from a single image. Comput & Graphics 115:339–347. https://doi.org/10.1016/j.cag.2023.07.011
Zhou K, Han X, Jiang N et al (2022) Hemlets posh: Learning part-centric heatmap triplets for 3d human pose and shape estimation. IEEE Trans Pattern Anal Machine Intell 44(6):3000–3014. https://doi.org/10.1109/TPAMI.2021.3051173
Chen D, Song Y, Liang F et al (2023) 3d human body reconstruction based on smpl model. Visual Comput 39(5):1893–1906. https://doi.org/10.1007/s00371-022-02453-x
Lu Y, Yu H, Ni W et al (2023) 3d real-time human reconstruction with a single rgbd camera. Appl Intell 53(8):8735–8745. https://doi.org/10.1007/s10489-022-03969-4
Khirodkar R, Tripathi S, Kitani K (2022) Occluded human mesh recovery. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 1715–1725
He K, Zhang X, Ren S et al (2016) Deep residual learning for image recognition. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 770–778
Sun K, Xiao B, Liu D et al (2019) Deep high-resolution representation learning for human pose estimation. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 5693–5703
Li Z, Oskarsson M, Heyden A (2022) Detailed 3d human body reconstruction from multi-view images combining voxel super-resolution and learned implicit representation. Appl Intell 52(6):6739–6759. https://doi.org/10.1007/s10489-021-02783-8
Xu W, Xiang D, Wang G et al (2022) Multiview video-based 3-d pose estimation of patients in computer-assisted rehabilitation environment (caren). IEEE Trans Human-Mach Syst 52(2):196–206. https://doi.org/10.1109/THMS.2022.3142108
Gerats BG, Wolterink JM, Broeders IA (2023) 3d human pose estimation in multi-view operating room videos using differentiable camera projections. Comput Methods Biomech Biomed Eng: Imaging & Visualization 11(4):1197–1205. https://doi.org/10.1080/21681163.2022.2155580
Shuai H, Wu L, Liu Q (2023) Adaptive multi-view and temporal fusing transformer for 3d human pose estimation. IEEE Trans Pattern Anal Machine Intell 45(4):4122–4135. https://doi.org/10.1109/TPAMI.2022.3188716
Zhou ZH (2018) A brief introduction to weakly supervised learning. National Sci Rev 5(1):44–53. https://doi.org/10.1093/NSR/NWX106
Zhou ZH, Li M (2010) Semi-supervised learning by disagreement. Knowl Inform Syst 24:415–439. https://doi.org/10.1007/s10115-009-0209-z
Eren ME, Bhattarai M, Joyce RJ et al (2023) Semi-supervised classification of malware families under extreme class imbalance via hierarchical non-negative matrix factorization with automatic model selection. ACM Trans Privacy Secur 26(4):1–27. https://doi.org/10.1145/3624567
Wu L, Fang L, He X et al (2023) Querying labeled for unlabeled: Cross-image semantic consistency guided semi-supervised semantic segmentation. IEEE Trans Pattern Anal Mach Intell 45(7):8827–8844. https://doi.org/10.1109/TPAMI.2022.3233584
Yang X, Song Z, King I et al (2023) A survey on deep semi-supervised learning. IEEE Trans Knowl Data Eng 35(9):8934–8954. https://doi.org/10.1109/TKDE.2022.3220219
Zhao H, Jia J, Koltun V (2020) Exploring self-attention for image recognition. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 10076–10085
Wenxuan Z, Yaqin Z, Zhaoxiang Z et al (2023) Lite transformer network with long-short range attention for real-time fire detection. Fire Technol 59(6):3231–3253. https://doi.org/10.1007/s10694-023-01465-w
Ranjan A, Bolkart T, Sanyal S et al (2018) Generating 3d faces using convolutional mesh autoencoders. In: Proceedings of the European conference on computer vision, Springer, pp 704–720
Pang S, Peng R, Dong Y et al (2023) Jointmetro: a 3d reconstruction model for human figures in works of art based on transformer. Neural Comput Appl pp 1–15. https://doi.org/10.1007/s00521-023-08844-y
Kocabas M, Huang CHP, Hilliges O et al (2021) Pare: Part attention regressor for 3d human body estimation. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 11127–11137
Ionescu C, Papava D, Olaru V et al (2013) Human3. 6m: Large scale datasets and predictive methods for 3d human sensing in natural environments. IEEE Trans Pattern Anal Machine Intell 36(7):1325–1339. https://doi.org/10.1109/TPAMI.2013.248
Andriluka M, Pishchulin L, Gehler P et al (2014) 2d human pose estimation: New benchmark and state of the art analysis. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 3686–3693
Mehta D, Sotnychenko O, Mueller F et al (2018) Single-shot multi-person 3d pose estimation from monocular rgb. In: Proceedings of the IEEE International Conference on 3D vision, pp 120–130
Lassner C, Romero J, Kiefel M et al (2017) Unite the people: Closing the loop between 3d and 2d human representations. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 6050–6059
Lin TY, Maire M, Belongie S et al (2014) Microsoft coco: Common objects in context. In: Proceedings of the European conference on computer vision, Springer, pp 740–755
Mehta D, Rhodin H, Casas D et al (2017) Monocular 3d human pose estimation in the wild using improved cnn supervision. In: Proceedings of the IEEE International Conference on 3D vision, pp 506–516
Deng J, Dong W, Socher R et al (2009) Imagenet: A large-scale hierarchical image database. In: Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp 248–255
Loshchilov I, Hutter F (2018) Decoupled weight decay regularization. In: Proceedings of the International Conference on Learning Representations, pp 1–18
Wang L, Liu X, Ma X et al (2022) A progressive quadric graph convolutional network for 3d human mesh recovery. IEEE Trans Circuits Syst Video Technol 33(1):104–117. https://doi.org/10.1109/TCSVT.2022.3199201
Kolotouros N, Pavlakos G, Jayaraman D et al (2021) Probabilistic modeling for human mesh recovery. In: Proceedings of the IEEE/CVF international conference on computer vision, pp 11605–11614
Yu Z, Zhang L, Xu Y et al (2022) Multiview human body reconstruction from uncalibrated cameras. In: Proceedings of the International Conference on neural information processing systems, pp 7879–7891
Funding
This work was supported in part by the Youth Innovation Promotion Association of Chinese Academy of Sciences (Y202072) and in part by the Natural Science Foundation of Shandong Province (ZR2021QE205).
Author information
Authors and Affiliations
Contributions
Conceptualization: Silong Sheng; Methodology: Silong Sheng, Tianyou Zheng, Zhijie Ren; Formal analysis and investigation: Silong Sheng, Tianyou Zheng; Writing - original draft preparation: Silong Sheng; Writing - review and editing: Tianyou Zheng, Weiwei Fu, Yang Zhang, Zhijie Ren; Funding acquisition: Weiwei Fu, Yang Zhang; Resources: Weiwei Fu, Yang Zhang; Supervision: Weiwei Fu.
Corresponding authors
Ethics declarations
Competing interests
The authors have no competing interests to declare that are relevant to the content of this article.
Ethics approval
Not applicable.
Consent to participate
Not applicable.
Consent for publication
Not applicable.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Sheng, S., Zheng, T., Ren, Z. et al. SS-MVMETRO: Semi-supervised multi-view human mesh recovery transformer. Appl Intell 54, 5027–5043 (2024). https://doi.org/10.1007/s10489-024-05435-9
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-024-05435-9