Reconstruction of 3D human body pose based on top-down learning

Hee Deok Yang, Sung K. Park, Seong Whan Lee

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

This paper presents a novel method for reconstructing 3D human body pose from monocular image sequences based on top-down learning. Human body pose is represented by a linear combination of prototypes of 2D silhouette images and their corresponding 3D body models in terms of the position of a predetermined set of joints. With a 2D silhouette image, we can estimate optimal coefficients for a linear combination of prototypes of the 2D silhouette images by solving least square minimization. The 3D body model of the input silhouette image is obtained by applying the estimated coefficients to the corresponding 3D body model of prototypes. In the learning stage, the proposed method is hierarchically constructed by classifying the training data into several clusters recursively. Also, in the reconstructing stage, the proposed method hierarchically reconstructs 3D human body pose with a silhouette image or a silhouette history image. We use a silhouette history image and a blurring silhouette image as the spatio-temporal features for reducing noise due to extraction of silhouette image and for extending the search area of current body pose to related body pose. The experimental results show that our method can be efficient and effective for reconstructing 3D human body pose.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science
EditorsD.-S. Huang, X.-P. Zhang, G.-B. Huang
Pages601-610
Number of pages10
Volume3644
EditionPART I
Publication statusPublished - 2005
EventInternational Conference on Intelligent Computing, ICIC 2005 - Hefei, China
Duration: 2005 Aug 232005 Aug 26

Other

OtherInternational Conference on Intelligent Computing, ICIC 2005
CountryChina
CityHefei
Period05/8/2305/8/26

ASJC Scopus subject areas

  • Computer Science (miscellaneous)

Cite this

Yang, H. D., Park, S. K., & Lee, S. W. (2005). Reconstruction of 3D human body pose based on top-down learning. In D-S. Huang, X-P. Zhang, & G-B. Huang (Eds.), Lecture Notes in Computer Science (PART I ed., Vol. 3644, pp. 601-610)

Reconstruction of 3D human body pose based on top-down learning. / Yang, Hee Deok; Park, Sung K.; Lee, Seong Whan.

Lecture Notes in Computer Science. ed. / D.-S. Huang; X.-P. Zhang; G.-B. Huang. Vol. 3644 PART I. ed. 2005. p. 601-610.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Yang, HD, Park, SK & Lee, SW 2005, Reconstruction of 3D human body pose based on top-down learning. in D-S Huang, X-P Zhang & G-B Huang (eds), Lecture Notes in Computer Science. PART I edn, vol. 3644, pp. 601-610, International Conference on Intelligent Computing, ICIC 2005, Hefei, China, 05/8/23.
Yang HD, Park SK, Lee SW. Reconstruction of 3D human body pose based on top-down learning. In Huang D-S, Zhang X-P, Huang G-B, editors, Lecture Notes in Computer Science. PART I ed. Vol. 3644. 2005. p. 601-610
Yang, Hee Deok ; Park, Sung K. ; Lee, Seong Whan. / Reconstruction of 3D human body pose based on top-down learning. Lecture Notes in Computer Science. editor / D.-S. Huang ; X.-P. Zhang ; G.-B. Huang. Vol. 3644 PART I. ed. 2005. pp. 601-610
@inproceedings{7624fbc03c4a42e4977c49c604a53eb1,
title = "Reconstruction of 3D human body pose based on top-down learning",
abstract = "This paper presents a novel method for reconstructing 3D human body pose from monocular image sequences based on top-down learning. Human body pose is represented by a linear combination of prototypes of 2D silhouette images and their corresponding 3D body models in terms of the position of a predetermined set of joints. With a 2D silhouette image, we can estimate optimal coefficients for a linear combination of prototypes of the 2D silhouette images by solving least square minimization. The 3D body model of the input silhouette image is obtained by applying the estimated coefficients to the corresponding 3D body model of prototypes. In the learning stage, the proposed method is hierarchically constructed by classifying the training data into several clusters recursively. Also, in the reconstructing stage, the proposed method hierarchically reconstructs 3D human body pose with a silhouette image or a silhouette history image. We use a silhouette history image and a blurring silhouette image as the spatio-temporal features for reducing noise due to extraction of silhouette image and for extending the search area of current body pose to related body pose. The experimental results show that our method can be efficient and effective for reconstructing 3D human body pose.",
author = "Yang, {Hee Deok} and Park, {Sung K.} and Lee, {Seong Whan}",
year = "2005",
language = "English",
volume = "3644",
pages = "601--610",
editor = "D.-S. Huang and X.-P. Zhang and G.-B. Huang",
booktitle = "Lecture Notes in Computer Science",
edition = "PART I",

}

TY - GEN

T1 - Reconstruction of 3D human body pose based on top-down learning

AU - Yang, Hee Deok

AU - Park, Sung K.

AU - Lee, Seong Whan

PY - 2005

Y1 - 2005

N2 - This paper presents a novel method for reconstructing 3D human body pose from monocular image sequences based on top-down learning. Human body pose is represented by a linear combination of prototypes of 2D silhouette images and their corresponding 3D body models in terms of the position of a predetermined set of joints. With a 2D silhouette image, we can estimate optimal coefficients for a linear combination of prototypes of the 2D silhouette images by solving least square minimization. The 3D body model of the input silhouette image is obtained by applying the estimated coefficients to the corresponding 3D body model of prototypes. In the learning stage, the proposed method is hierarchically constructed by classifying the training data into several clusters recursively. Also, in the reconstructing stage, the proposed method hierarchically reconstructs 3D human body pose with a silhouette image or a silhouette history image. We use a silhouette history image and a blurring silhouette image as the spatio-temporal features for reducing noise due to extraction of silhouette image and for extending the search area of current body pose to related body pose. The experimental results show that our method can be efficient and effective for reconstructing 3D human body pose.

AB - This paper presents a novel method for reconstructing 3D human body pose from monocular image sequences based on top-down learning. Human body pose is represented by a linear combination of prototypes of 2D silhouette images and their corresponding 3D body models in terms of the position of a predetermined set of joints. With a 2D silhouette image, we can estimate optimal coefficients for a linear combination of prototypes of the 2D silhouette images by solving least square minimization. The 3D body model of the input silhouette image is obtained by applying the estimated coefficients to the corresponding 3D body model of prototypes. In the learning stage, the proposed method is hierarchically constructed by classifying the training data into several clusters recursively. Also, in the reconstructing stage, the proposed method hierarchically reconstructs 3D human body pose with a silhouette image or a silhouette history image. We use a silhouette history image and a blurring silhouette image as the spatio-temporal features for reducing noise due to extraction of silhouette image and for extending the search area of current body pose to related body pose. The experimental results show that our method can be efficient and effective for reconstructing 3D human body pose.

UR - http://www.scopus.com/inward/record.url?scp=27144465455&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=27144465455&partnerID=8YFLogxK

M3 - Conference contribution

VL - 3644

SP - 601

EP - 610

BT - Lecture Notes in Computer Science

A2 - Huang, D.-S.

A2 - Zhang, X.-P.

A2 - Huang, G.-B.

ER -