Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Template based human pose and shape estimation from a single RGB-D image

Li, Zhongguo LU ; Heyden, Anders LU orcid and Oskarsson, Magnus LU orcid (2019) 8th International Conference on Pattern Recognition Applications and Methods, ICPRAM 2019 p.574-581
Abstract

Estimating the 3D model of the human body is needed for many applications. However, this is a challenging problem since the human body inherently has a high complexity due to self-occlusions and articulation. We present a method to reconstruct the 3D human body model from a single RGB-D image. 2D joint points are firstly predicted by a CNN-based model called convolutional pose machine, and the 3D joint points are calculated using the depth image. Then, we propose to utilize both 2D and 3D joint points, which provide more information, to fit a parametric body model (SMPL). This is implemented through minimizing an objective function, which measures the difference of the joint points between the observed model and the parametric model.... (More)

Estimating the 3D model of the human body is needed for many applications. However, this is a challenging problem since the human body inherently has a high complexity due to self-occlusions and articulation. We present a method to reconstruct the 3D human body model from a single RGB-D image. 2D joint points are firstly predicted by a CNN-based model called convolutional pose machine, and the 3D joint points are calculated using the depth image. Then, we propose to utilize both 2D and 3D joint points, which provide more information, to fit a parametric body model (SMPL). This is implemented through minimizing an objective function, which measures the difference of the joint points between the observed model and the parametric model. The pose and shape parameters of the body are obtained through optimization and the final 3D model is estimated. The experiments on synthetic data and real data demonstrate that our method can estimate the 3D human body model correctly.

(Less)
Please use this url to cite or link to this publication:
author
; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
2D and 3D Pose, Human Body Reconstruction, Pose, Shape Estimation, SMPL Model
host publication
ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods
editor
Fred, Ana ; De Marsico, Maria and di Baja, Gabriella Sanniti
pages
8 pages
publisher
SciTePress
conference name
8th International Conference on Pattern Recognition Applications and Methods, ICPRAM 2019
conference location
Prague, Czech Republic
conference dates
2019-02-19 - 2019-02-21
external identifiers
  • scopus:85064634552
ISBN
9789897583513
DOI
10.5220/0007383605740581
language
English
LU publication?
yes
id
3c23bf11-b858-4f3d-8b4c-230aad74ee7c
date added to LUP
2019-05-07 09:08:43
date last changed
2023-12-03 07:46:22
@inproceedings{3c23bf11-b858-4f3d-8b4c-230aad74ee7c,
  abstract     = {{<p>Estimating the 3D model of the human body is needed for many applications. However, this is a challenging problem since the human body inherently has a high complexity due to self-occlusions and articulation. We present a method to reconstruct the 3D human body model from a single RGB-D image. 2D joint points are firstly predicted by a CNN-based model called convolutional pose machine, and the 3D joint points are calculated using the depth image. Then, we propose to utilize both 2D and 3D joint points, which provide more information, to fit a parametric body model (SMPL). This is implemented through minimizing an objective function, which measures the difference of the joint points between the observed model and the parametric model. The pose and shape parameters of the body are obtained through optimization and the final 3D model is estimated. The experiments on synthetic data and real data demonstrate that our method can estimate the 3D human body model correctly.</p>}},
  author       = {{Li, Zhongguo and Heyden, Anders and Oskarsson, Magnus}},
  booktitle    = {{ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods}},
  editor       = {{Fred, Ana and De Marsico, Maria and di Baja, Gabriella Sanniti}},
  isbn         = {{9789897583513}},
  keywords     = {{2D and 3D Pose; Human Body Reconstruction; Pose; Shape Estimation; SMPL Model}},
  language     = {{eng}},
  pages        = {{574--581}},
  publisher    = {{SciTePress}},
  title        = {{Template based human pose and shape estimation from a single RGB-D image}},
  url          = {{http://dx.doi.org/10.5220/0007383605740581}},
  doi          = {{10.5220/0007383605740581}},
  year         = {{2019}},
}