南京师范大学计算机与电子信息/人工智能学院,南京 210023
吉根林,教授,博士生导师,E-mail:glji@njnu.edu.cn。
收稿:2025-07-05,
修回:2025-10-24,
纸质出版:2026-04-28
移动端阅览
朱炜,吉根林,王泽宇,等. 基于姿态约束的3D人体网格细化算法[J]. 南京航空航天大学学报(自然科学版),2026,58(2):449⁃456.
ZHU Wei, JI Genlin, WANG Zeyu, et al. Pose‑constrained 3D human mesh refinement algorithm[J]. Journal of Nanjing University of Aeronautics & Astronautics(Natural Science Edition),2026, 58(2):449⁃456.
朱炜,吉根林,王泽宇,等. 基于姿态约束的3D人体网格细化算法[J]. 南京航空航天大学学报(自然科学版),2026,58(2):449⁃456. DOI: 10.16356/j.2097-6771.2026.02.021.
ZHU Wei, JI Genlin, WANG Zeyu, et al. Pose‑constrained 3D human mesh refinement algorithm[J]. Journal of Nanjing University of Aeronautics & Astronautics(Natural Science Edition),2026, 58(2):449⁃456. DOI: 10.16356/j.2097-6771.2026.02.021.
3D人体网格恢复(Human mesh recovery,HMR)在虚拟现实、增强现实、自动驾驶和运动科学等领域具有广泛应用前景。然而,现有方法面临着3D模型与2D关键点无法精确对齐的问题,且在恢复过程中常忽视人体固有的对称性和比例约束,导致恢复的3D人体网格模型不符合人体基本特征。本文提出了一种基于姿态约束的人体网格细化(Pose‑constrained human mesh refinement,PC‑HMR)算法,通过多维约束融合策略优化形状参数和姿态参数。在形状参数优化方面,引入人体对称性特征和骨骼比例约束;在姿态参数优化方面,设计了一种多算法融合的姿态约束建模框架,通过系统性整合多种互补算法来替代传统的乘法路径选择机制。为了评估本文方法的有效性,在3DPW和Human3.6M两个标准数据集上进行了实验。实验结果表明,相比现有最优方法,PC‑HMR综合性能更优。消融实验进一步验证了各模块对算法性能的影响。
3D human mesh recovery (HMR) has extensive applications in virtual reality, augmented reality, autonomous driving, and sports science. However, the existing methods suffer from imprecise alignment between 3D models and 2D keypoints and often neglect the inherent symmetry and proportion constraints of the human body, resulting in recovered meshes that violate fundamental human characteristics.Our paper proposes a pose-constrained human mesh refinement (PC-HMR) algorithm that optimizes shape and pose parameters through constraint fusion. For shape parameter optimization, we introduce human body symmetry features and skeletal proportion constraints. For pose parameter optimization, we designe a multi-algorithm fusion framework that systematically integrates complementary algorithms to replace traditional single-path selection mechanisms.We evaluate the proposed method in two standard datasets: 3DPW and Human3.6M. Experimental results demonstrate that PC-HMR achieve superior performance compared to existing state-of-the-art methods. Ablation studies further validate the contribution of each module to overall performance.
LIU Yang , QIU Changzhen , ZHANG Zhiyong . Deep learning for 3D human pose estimation and mesh recovery: A survey [J]. Neurocomputing , 2024 , 596 : 128049 .
TIAN Yating , ZHANG Hongwen , LIU Yebin , et al . Recovering 3D human mesh from monocular images: A survey [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence , 2023 , 45 ( 12 ): 15406 - 15425 .
CHO H , CHO Y , AHN J , et al . Implicit 3D human mesh recovery using consistency with pose and shape from unseen-view [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Vancouver, Canada : [s.n.] , 2023 : 21148 - 21158 .
KANAZAWA A , BLACK M J , JACOBS D W , et al . End-to-end recovery of human shape and pose [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Salt Lake City, USA : [s.n.] , 2018 : 7122 - 7131 .
LIN Kevin , WANG Lijuan , LIU Zicheng . Mesh graphormer [C]// Proceedings of the IEEE/CVF International Conference on Computer Vision , Virtual Event, 2021 . Piscataway, NJ, USA : Institute of Electrical and Electronics Engineers , 2021 : 12939 - 12948 .
LIN Kevin , WANG Lijuan , LIU Zicheng . End-to-end human pose and mesh reconstruction with transformers [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Virtual Event , 2021 . Piscataway, NJ, USA : Institute of Electrical and Electronics Engineers , 2021: 1954 - 1963 .
LOPER M , MAHMOOD N , ROMEROJ , et al . SMPL: A skinned multi-person linear model [C]// Proceedings of Seminal Graphics Papers: Pushing the Boundaries , Volume 2 . New York, NY, USA : Association for Computing Machinery , 2023: 851 - 866 .
IONESCU C , PAPAVA D , OLARU V , et al . Human3.6m: Large scale datasets and predictive methods for 3d human sensing in natural environments [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence , 2013 , 36 ( 7 ): 1325 - 1339 .
VON MARCARD T , HENSCHEL R , BLACK M J , et al . Recovering accurate 3d human pose in the wild using imus and a moving camera [C]// Proceedings of the European Conference on Computer Vision . Munich, Germany : [s.n.] , 2018 : 601 - 617 .
DOU Zhiyang , WU Qingxuan , LIN Cheng , et al . Tore: Token reduction for efficient human mesh recovery with transformer [C]// Proceedings of the IEEE/CVF International Conference on Computer Vision . Paris, France : [s.n.] , 2023 : 15143 - 15155 .
WANG Dongkai , ZHANG Shiliang . 3D human mesh recovery with sequentially global rotation estimation [C]// Proceedings of the IEEE/CVF International Conference on Computer Vision . Paris, France : [s.n.] , 2023 : 14953 - 14962 .
SHEN Xiaolong , YANG Zongxin , WANG Xiaohan , et al . Global-to-local modeling for video-based 3D human pose and shape estimation [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Vancouver, Canada : [s.n.] , 2023 : 8887 - 8896 .
CHOI H , MOON G , CHANG J Y , et al . Beyond static features for temporally consistent 3Dhuman pose and shape from a video [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Virtual Event , 2021 . Piscataway, NJ, USA : Institute of Electrical and Electronics Engineers , 2021: 1964 - 1973 .
WEI Wenli , LIN Jenchun , LIU Tyngluh , et al . Capturing humans in motion: Temporal-attentive 3D human pose and shape estimation from monocular video [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . New Orleans, USA : [s.n.] , 2022 : 13211 - 13220 .
LI Jiefeng , XU Chao , CHEN Zhicun , et al . Hybrik: A hybrid analytical-neural inverse kinematics solution for 3D human pose and shape estimation [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Virtual Event , 2021 . Piscataway, NJ, USA : Institute of Electrical and Electronics Engineers , 2021: 3383 - 3393 .
BOGO F , KANAZAWA A , LASSNER C , et al . Keep it SMPL: Automatic estimation of 3d human pose and shape from a single image [C]// Proceedings of the European Conference on Computer Vision . Amsterdam, The Netherlands : [s.n.] , 2016 : 561 - 578 .
KOLOTOUROS N , PAVLAKOS G , BLACK M J , et al . Learning to reconstruct 3D human pose and shape via model-fitting in the loop [C]// Proceedings of the IEEE/CVF International Conference on Computer Vision . Seoul, South Korea : [s.n.] , 2019 : 2252 - 2261 .
GUAN P , WEISS A , BALAN A O , et al . Estimating human shape and pose from a single image [C]// Proceedings of the IEEE International Conference on Computer Vision . Kyoto, Japan : [s.n.] , 2009 : 1381 - 1388 .
JOOH , NEVEROVA N , VEDALDI A . Exemplar fine-tuning for 3D human model fitting towards in-the-wild 3D human pose estimation [C]// Proceedings of the 2021 International Conference on 3D Vision . Virtual Event , 2021 . Piscataway, NJ, USA : Institute of Electrical and Electronics Engineers , 2021: 42 - 52 .
LI Zhihao , LIU Jianzhuang , ZHANG Zhensong , et al . Cliff: Carrying location information in full frames into human pose and shape estimation [C]// Proceedings of the European Conference on Computer Vision . Tel Aviv, Israel : [s.n.] , 2022 : 590 - 606 .
YANG Fengyuan , GU Kerui , YAO A . KITRO: Refining human mesh by 2D clues and kinematic-tree rotation [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Seattle, USA : [s.n.] , 2024 : 1052 - 1061 .
BOGIN B , VARELA-SILVA M I . Leg length, body proportion, and health: A review with a note on beauty [J]. International journal of environmental research and public health , 2010 , 7 ( 3 ): 1047 - 1075 .
WANG Nan , MA Jie , JIN Dan , et al . A special golden curve in human upper limbs’ length proportion: A functional partition which is different from anatomy [J]. BioMed Research International , 2017 :Article No. 4158561 .
0
浏览量
0
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010602201714号