Biased manifold learning for view invariant body pose estimation

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In human body pose estimation, manifold learning has been considered as a useful method with regard to reducing the dimension of 2D images and 3D body configuration data. Most commonly, body pose is estimated from silhouettes derived from images or image sequences. A major problem in applying manifold estimation to pose estimation is its vulnerability to silhouette variation caused by changes of factors such as viewpoint, person, and distance. In this paper, we propose a novel approach that combines three separate manifolds for viewpoint, pose, and 3D body configuration focusing on the problem of viewpoint-induced silhouette variation. The biased manifold learning is used to learn these manifolds with appropriately weighted distances. The proposed method requires four mapping functions that are learned by a generalized regression neural network for robustness. Despite the use of only three manifolds, experimental results show that the proposed method can reliably estimate 3D body poses from 2D images with all learned viewpoints.

Original languageEnglish
Article number1250058
JournalInternational Journal of Wavelets, Multiresolution and Information Processing
Volume10
Issue number6
DOIs
Publication statusPublished - 2012 Nov 1

Fingerprint

Manifold Learning
Pose Estimation
Silhouette
Biased
Three-manifolds
Invariant
Configuration
Image Sequence
Vulnerability
Person
Regression
Neural Networks
Robustness
Neural networks
Experimental Results
Estimate

Keywords

  • 3D pose estimation
  • manifold learning
  • nonlinear dimensionality reduction

ASJC Scopus subject areas

  • Applied Mathematics
  • Information Systems
  • Signal Processing

Cite this

@article{5b1964f9f6aa4771b2ed3c29dd0c5c66,
title = "Biased manifold learning for view invariant body pose estimation",
abstract = "In human body pose estimation, manifold learning has been considered as a useful method with regard to reducing the dimension of 2D images and 3D body configuration data. Most commonly, body pose is estimated from silhouettes derived from images or image sequences. A major problem in applying manifold estimation to pose estimation is its vulnerability to silhouette variation caused by changes of factors such as viewpoint, person, and distance. In this paper, we propose a novel approach that combines three separate manifolds for viewpoint, pose, and 3D body configuration focusing on the problem of viewpoint-induced silhouette variation. The biased manifold learning is used to learn these manifolds with appropriately weighted distances. The proposed method requires four mapping functions that are learned by a generalized regression neural network for robustness. Despite the use of only three manifolds, experimental results show that the proposed method can reliably estimate 3D body poses from 2D images with all learned viewpoints.",
keywords = "3D pose estimation, manifold learning, nonlinear dimensionality reduction",
author = "Dongcheol Hur and Heung-Il Suk and Christian Wallraven and Lee, {Seong Whan}",
year = "2012",
month = "11",
day = "1",
doi = "10.1142/S0219691312500580",
language = "English",
volume = "10",
journal = "International Journal of Wavelets, Multiresolution and Information Processing",
issn = "0219-6913",
publisher = "World Scientific Publishing Co. Pte Ltd",
number = "6",

}

TY - JOUR

T1 - Biased manifold learning for view invariant body pose estimation

AU - Hur, Dongcheol

AU - Suk, Heung-Il

AU - Wallraven, Christian

AU - Lee, Seong Whan

PY - 2012/11/1

Y1 - 2012/11/1

N2 - In human body pose estimation, manifold learning has been considered as a useful method with regard to reducing the dimension of 2D images and 3D body configuration data. Most commonly, body pose is estimated from silhouettes derived from images or image sequences. A major problem in applying manifold estimation to pose estimation is its vulnerability to silhouette variation caused by changes of factors such as viewpoint, person, and distance. In this paper, we propose a novel approach that combines three separate manifolds for viewpoint, pose, and 3D body configuration focusing on the problem of viewpoint-induced silhouette variation. The biased manifold learning is used to learn these manifolds with appropriately weighted distances. The proposed method requires four mapping functions that are learned by a generalized regression neural network for robustness. Despite the use of only three manifolds, experimental results show that the proposed method can reliably estimate 3D body poses from 2D images with all learned viewpoints.

AB - In human body pose estimation, manifold learning has been considered as a useful method with regard to reducing the dimension of 2D images and 3D body configuration data. Most commonly, body pose is estimated from silhouettes derived from images or image sequences. A major problem in applying manifold estimation to pose estimation is its vulnerability to silhouette variation caused by changes of factors such as viewpoint, person, and distance. In this paper, we propose a novel approach that combines three separate manifolds for viewpoint, pose, and 3D body configuration focusing on the problem of viewpoint-induced silhouette variation. The biased manifold learning is used to learn these manifolds with appropriately weighted distances. The proposed method requires four mapping functions that are learned by a generalized regression neural network for robustness. Despite the use of only three manifolds, experimental results show that the proposed method can reliably estimate 3D body poses from 2D images with all learned viewpoints.

KW - 3D pose estimation

KW - manifold learning

KW - nonlinear dimensionality reduction

UR - http://www.scopus.com/inward/record.url?scp=84871588370&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871588370&partnerID=8YFLogxK

U2 - 10.1142/S0219691312500580

DO - 10.1142/S0219691312500580

M3 - Article

AN - SCOPUS:84871588370

VL - 10

JO - International Journal of Wavelets, Multiresolution and Information Processing

JF - International Journal of Wavelets, Multiresolution and Information Processing

SN - 0219-6913

IS - 6

M1 - 1250058

ER -