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

Forskningsoutput: Kapitel i bok/rapport/Conference proceedingKonferenspaper i proceeding

Standard

Template based human pose and shape estimation from a single RGB-D image. / Li, Zhongguo; Heyden, Anders; Oskarsson, Magnus.

ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods. red. / Ana Fred; Maria De Marsico; Gabriella Sanniti di Baja. SciTePress, 2019. s. 574-581.

Forskningsoutput: Kapitel i bok/rapport/Conference proceedingKonferenspaper i proceeding

Harvard

Li, Z, Heyden, A & Oskarsson, M 2019, Template based human pose and shape estimation from a single RGB-D image. i A Fred, M De Marsico & GS di Baja (red), ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods. SciTePress, s. 574-581, 8th International Conference on Pattern Recognition Applications and Methods, ICPRAM 2019, Prague, Tjeckien, 2019/02/19. https://doi.org/10.5220/0007383605740581

APA

Li, Z., Heyden, A., & Oskarsson, M. (2019). Template based human pose and shape estimation from a single RGB-D image. I A. Fred, M. De Marsico, & G. S. di Baja (Red.), ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods (s. 574-581). SciTePress. https://doi.org/10.5220/0007383605740581

CBE

Li Z, Heyden A, Oskarsson M. 2019. Template based human pose and shape estimation from a single RGB-D image. Fred A, De Marsico M, di Baja GS, redaktörer. I ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods. SciTePress. s. 574-581. https://doi.org/10.5220/0007383605740581

MLA

Li, Zhongguo, Anders Heyden, och Magnus Oskarsson "Template based human pose and shape estimation from a single RGB-D image"., Fred, Ana De Marsico, Maria di Baja, Gabriella Sanniti (redaktörer). ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods. SciTePress. 2019, 574-581. https://doi.org/10.5220/0007383605740581

Vancouver

Li Z, Heyden A, Oskarsson M. Template based human pose and shape estimation from a single RGB-D image. I Fred A, De Marsico M, di Baja GS, redaktörer, ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods. SciTePress. 2019. s. 574-581 https://doi.org/10.5220/0007383605740581

Author

Li, Zhongguo ; Heyden, Anders ; Oskarsson, Magnus. / Template based human pose and shape estimation from a single RGB-D image. ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods. redaktör / Ana Fred ; Maria De Marsico ; Gabriella Sanniti di Baja. SciTePress, 2019. s. 574-581

RIS

TY - GEN

T1 - Template based human pose and shape estimation from a single RGB-D image

AU - Li, Zhongguo

AU - Heyden, Anders

AU - Oskarsson, Magnus

PY - 2019

Y1 - 2019

N2 - 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.

AB - 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.

KW - 2D and 3D Pose

KW - Human Body Reconstruction

KW - Pose

KW - Shape Estimation

KW - SMPL Model

U2 - 10.5220/0007383605740581

DO - 10.5220/0007383605740581

M3 - Paper in conference proceeding

AN - SCOPUS:85064634552

SP - 574

EP - 581

BT - ICPRAM 2019 - Proceedings of the 8th International Conference on Pattern Recognition Applications and Methods

A2 - Fred, Ana

A2 - De Marsico, Maria

A2 - di Baja, Gabriella Sanniti

PB - SciTePress

T2 - 8th International Conference on Pattern Recognition Applications and Methods, ICPRAM 2019

Y2 - 19 February 2019 through 21 February 2019

ER -