Filters








6,578 Hits in 5.1 sec

3D Human Mesh Regression with Dense Correspondence [article]

Wang Zeng, Wanli Ouyang, Ping Luo, Wentao Liu, Xiaogang Wang
2021 arXiv   pre-print
Finally we reconstruct 3D human mesh from the regressed location map with a predefined mapping function.  ...  This paper proposes a model-free 3D human mesh estimation framework, named DecoMR, which explicitly establishes the dense correspondence between the mesh and the local image features in the UV space (i.e  ...  Similarly, we show the results on SURREAL dataset in 3D Human Mesh Regression with Dense Correspondence **Supplementary Material** Wang Zeng 1 , Wanli Ouyang 2 , Ping Luo 3 , Wentao Liu 4 , and Xiaogang  ... 
arXiv:2006.05734v2 fatcat:nm5ygdcsp5fbnfjn2zxepavrde

Monocular Human Shape and Pose with Dense Mesh-borne Local Image Features [article]

Shubhendu Jena, Franck Multon, Adnane Boukhayma
2021 arXiv   pre-print
These features are sampled from the CNN image feature maps by utilizing pixel-to-mesh correspondences generated with DensePose.  ...  feature appended to all mesh vertices equally to initialize the GCN stage, which transforms a template T-posed mesh into the target pose.  ...  dense correspondence. [37] , [30] utilized DensePose correspondence maps for 3D human model recovery.  ... 
arXiv:2111.05319v3 fatcat:q3o3w3cxtfglnbyb7tvfpe62xi

PyMAF: 3D Human Pose and Shape Regression with Pyramidal Mesh Alignment Feedback Loop [article]

Hongwen Zhang, Yating Tian, Xinchi Zhou, Wanli Ouyang, Yebin Liu, Limin Wang, Zhenan Sun
2021 arXiv   pre-print
Regression-based methods have recently shown promising results in reconstructing human meshes from monocular images.  ...  To reduce noise and enhance the reliability of these evidences, an auxiliary pixel-wise supervision is imposed on the feature encoder, which provides mesh-image correspondence guidance for our network  ...  Human3.6M [17] is commonly used as the benchmark dataset for 3D human pose estimation, consisting of 3.6 million video frames captured in the controlled environment.  ... 
arXiv:2103.16507v4 fatcat:mlew2hk7dbhvpexr7t76rjxe4a

The Best of Both Worlds: Combining Model-based and Nonparametric Approaches for 3D Human Body Estimation [article]

Zhe Wang, Jimei Yang, Charless Fowlkes
2022 arXiv   pre-print
A dense map prediction module explicitly establishes the dense UV correspondence between the image evidence and each part of the body model.  ...  Experiments demonstrate that our framework outperforms existing 3D human estimation methods on multiple public benchmarks.  ...  The dense correspondence methods [46] [47] [48] , which are based on template SMPL human mesh surface and have been proven for various tasks.  ... 
arXiv:2205.00508v1 fatcat:2wxgajvk4zhifj4jiqtgt353ky

Convolutional Mesh Regression for Single-Image Human Shape Reconstruction [article]

Nikos Kolotouros, Georgios Pavlakos, Kostas Daniilidis
2019 arXiv   pre-print
Previous approaches consider a parametric model of the human body, SMPL, and attempt to regress the model parameters that give rise to a mesh consistent with image evidence.  ...  This paper addresses the problem of 3D human pose and shape estimation from a single image.  ...  In contrast to them, we retain the SMPL mesh topology, which allows us to get dense semantic correspondences of our 3D prediction with the image, and in the end we can also easily regress the model's parameters  ... 
arXiv:1905.03244v1 fatcat:cf7wrt4if5gd5c4rv53uon5zpq

Convolutional Mesh Regression for Single-Image Human Shape Reconstruction

Nikos Kolotouros, Georgios Pavlakos, Kostas Daniilidis
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Previous approaches consider a parametric model of the human body, SMPL, and attempt to regress the model parameters that give rise to a mesh consistent with image evidence.  ...  This paper addresses the problem of 3D human pose and shape estimation from a single image.  ...  In contrast to them, we retain the SMPL mesh topology, which allows us to get dense semantic correspondences of our 3D prediction with the image, and in the end we can also easily regress the model's parameters  ... 
doi:10.1109/cvpr.2019.00463 dblp:conf/cvpr/KolotourosPD19 fatcat:exyr5zm7dvdbrdge6vbeqzehme

HoloPose: Holistic 3D Human Reconstruction In-The-Wild

Riza Alp Guler, Iasonas Kokkinos
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
We introduce HoloPose, a method for holistic monocular 3D human body reconstruction.  ...  We further train a multi-task network comprising 2D, 3D and Dense Pose estimation to drive the 3D reconstruction task.  ...  Surface Correspondence Performance Dense correspondence measures 3D mesh alignment accuracy in challenging, "in the wild" scenarios.  ... 
doi:10.1109/cvpr.2019.01114 dblp:conf/cvpr/GulerK19 fatcat:z27zxf6qkjh5vdu3ogvuykgc7e

The use of pseudo-landmarks for craniofacial analysis: A comparative study with L1-regularized logistic regression

Ezgi Mercan, Linda G. Shapiro, Seth M. Weinberg, Su-In Lee
2013 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)  
We have developed a computationally simple method that can compute pseudo-landmark points at different resolutions from 3D meshes of human faces.  ...  In this paper, we perform a comparative study employing L1-regularized logistic regression to train a classifier that predicts the sex of 500 normal adult face meshes in order to compare our method to  ...  One project, the 3D Facial Norms Database, is studying the variation of facial forms in humans by constructing a repository of 3D facial data (in the form of 3D meshes) and genetic data.  ... 
doi:10.1109/embc.2013.6610940 pmid:24111127 pmcid:PMC3836675 dblp:conf/embc/MercanSWL13 fatcat:trjzj4fw3fb6ppbzawvlct3rmi

Learning Body Shape and Pose from Dense Correspondences

Yusuke Yoshiyasu, Lucas Gamez
2020 Annual Conference of the European Association for Computer Graphics  
The idea is to use dense correspondences between image points and a body surface, which can be annotated on in-the-wild 2D images, to extract, aggregate and learn 3D information such as body shape and  ...  In this paper, we address the problem of learning 3D human pose and body shape from 2D image dataset, without having to use 3D supervisions (body shape and pose) which are in practice difficult to obtain  ...  geometric constraints for 3D human pose estimation.  ... 
doi:10.2312/egs.20201012 dblp:conf/eurographics/YoshiyasuG20 fatcat:ffbrncstkvauzj4clmyi76paem

Dual Grid Net: hand mesh vertex regression from single depth maps [article]

Chengde Wan, Thomas Probst, Luc Van Gool, Angela Yao
2019 arXiv   pre-print
We present a method for recovering the dense 3D surface of the hand by regressing the vertex coordinates of a mesh model from a single depth map.  ...  With multi-camera rig during training to resolve self-occlusion, it can perform competitively with strongly supervised methods Without any human annotation.  ...  We then map features from the image grid to the mesh grid according to dense correspondence map and recover the 3D coordinates of all the mesh vertices(sec. 3.3).  ... 
arXiv:1907.10695v1 fatcat:67ln42zlujf73kgiwddedplruu

Analyzing Clothing Layer Deformation Statistics of 3D Human Motions [chapter]

Jinlong Yang, Jean-Sébastien Franco, Franck Hétroy-Wheeler, Stefanie Wuhrer
2018 Lecture Notes in Computer Science  
parameter regression model using neural networks on the other hand, designed to regress from any semantic parameter whose variation is observed in a training set, to the layer parameterization space.  ...  Recent capture technologies and methods allow not only to retrieve 3D model sequence of moving people in clothing, but also to separate and extract the underlying body geometry, motion component and the  ...  Processing raw dense 3D motion captures for human modeling.  ... 
doi:10.1007/978-3-030-01234-2_15 fatcat:dtm6ehlnrzb2xnxjq2snhfsbpm

DenseRaC: Joint 3D Pose and Shape Estimation by Dense Render-and-Compare [article]

Yuanlu Xu, Song-Chun Zhu, Tony Tung
2019 arXiv   pre-print
The generated data covers diversified camera views, human actions and body shapes, and is paired with full ground truth.  ...  Our two-step framework takes the body pixel-to-surface correspondence map (i.e., IUV map) as proxy representation and then performs estimation of parameterized human pose and shape.  ...  We would like to thank Tengyu Liu and Elan Markowitz for helping with data collection, Tuur Jan M Stuyck and Aaron Ferguson for cloth simulation, Natalia Neverova and colleagues at FRL, FAIR and UCLA for  ... 
arXiv:1910.00116v2 fatcat:u67kndqjqneufc6rkatpgshkhq

DenseRaC: Joint 3D Pose and Shape Estimation by Dense Render-and-Compare

Yuanlu Xu, Song-Chun Zhu, Tony Tung
2019 2019 IEEE/CVF International Conference on Computer Vision (ICCV)  
The generated data covers diversified camera views, human actions and body shapes, and is paired with full ground truth.  ...  Our two-step framework takes the body pixel-to-surface correspondence map (i.e., IUV map) as proxy representation and then performs estimation of parameterized human pose and shape.  ...  We would like to thank Tengyu Liu and Elan Markowitz for helping with data collection, Tuur Jan M Stuyck and Aaron Ferguson for cloth simulation, Natalia Neverova and colleagues at FRL, FAIR and UCLA for  ... 
doi:10.1109/iccv.2019.00785 dblp:conf/iccv/XuZT19 fatcat:u3c3lz62ezcyhev3fcbooo6lem

DensePose 3D: Lifting Canonical Surface Maps of Articulated Objects to the Third Dimension [article]

Roman Shapovalov, David Novotny, Benjamin Graham, Patrick Labatut, Andrea Vedaldi
2021 arXiv   pre-print
The method learns, in an end-to-end fashion, a soft partition of a given category-specific 3D template mesh into rigid parts together with a monocular reconstruction network that predicts the part motions  ...  We tackle the problem of monocular 3D reconstruction of articulated objects like humans and animals.  ...  It introduced a dataset with manually labelled correspondences as well as a new deep network architecture to regress dense correspondences from images.  ... 
arXiv:2109.00033v1 fatcat:mlnrvprdhjf5vdvn47wfae7ule

Topologically Consistent Multi-View Face Inference Using Volumetric Sampling [article]

Tianye Li and Shichen Liu and Timo Bolkart and Jiayi Liu and Hao Li and Yajie Zhao
2021 arXiv   pre-print
High-fidelity face digitization solutions often combine multi-view stereo (MVS) techniques for 3D reconstruction and a non-rigid registration step to establish dense correspondence across identities and  ...  We demonstrate state-of-the-art geometric and correspondence accuracy, while only taking 0.385 seconds to compute a mesh with 10K vertices, which is three orders of magnitude faster than traditional techniques  ...  duces meshes in dense vertex correspondence. rectly.  ... 
arXiv:2110.02948v1 fatcat:d6be6b66jneotjjco4ybkj2lym
« Previous Showing results 1 — 15 out of 6,578 results