Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets

Ronald Walter Poppe

Research output: Book/Report › Report › Professional

26 Downloads (Pure)

Abstract

We present an example-based approach to pose recovery, using histograms of oriented gradients as image descriptors. Tests on the HumanEva-I and HumanEva-II data sets provide us insight into the strengths and limitations of an example-based approach. We report mean relative 3D errors of approximately 65 mm per joint on HumanEva-I, and 175 mm on HumanEva-II. We discuss our results using single and multiple views. Also, we perform experiments to assess the algorithm’s generalization to unseen subjects, actions and viewpoints. We plan to incorporate the temporal aspect of human motion analysis to reduce orientation ambiguities, and increase the pose recovery accuracy.
Original languageUndefined
Place of PublicationEnschede
PublisherCentre for Telematics and Information Technology (CTIT)
Number of pages9
Publication statusPublished - Oct 2007

Publication series

NameCTIT Technical Report Series
PublisherUniversity of Twente, Centre for Telematics and Information Technology
No.LNCS4549/TR-CTIT-07-72
ISSN (Print)1381-3625

Keywords

  • METIS-242003
  • EC Grant Agreement nr.: FP6/033812
  • IR-64422
  • EWI-11258
  • HMI-CI: Computational Intelligence
  • HMI-HF: Human Factors
  • HMI-VRG: Virtual Reality and Graphics

Cite this

Poppe, R. W. (2007). Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets. (CTIT Technical Report Series; No. LNCS4549/TR-CTIT-07-72). Enschede: Centre for Telematics and Information Technology (CTIT).
Poppe, Ronald Walter. / Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets. Enschede : Centre for Telematics and Information Technology (CTIT), 2007. 9 p. (CTIT Technical Report Series; LNCS4549/TR-CTIT-07-72).
@book{730b94fd6bd1433097f1fe57db229fb9,
title = "Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets",
abstract = "We present an example-based approach to pose recovery, using histograms of oriented gradients as image descriptors. Tests on the HumanEva-I and HumanEva-II data sets provide us insight into the strengths and limitations of an example-based approach. We report mean relative 3D errors of approximately 65 mm per joint on HumanEva-I, and 175 mm on HumanEva-II. We discuss our results using single and multiple views. Also, we perform experiments to assess the algorithm’s generalization to unseen subjects, actions and viewpoints. We plan to incorporate the temporal aspect of human motion analysis to reduce orientation ambiguities, and increase the pose recovery accuracy.",
keywords = "METIS-242003, EC Grant Agreement nr.: FP6/033812, IR-64422, EWI-11258, HMI-CI: Computational Intelligence, HMI-HF: Human Factors, HMI-VRG: Virtual Reality and Graphics",
author = "Poppe, {Ronald Walter}",
note = "This text has been accepted for the Computer Vision and Pattern Recognition (CVPR{\^a}€™07) workshop on Evaluation of Articulated Human Motion and Pose Estimation EHuM2). This version of the document is for documentation purposes.",
year = "2007",
month = "10",
language = "Undefined",
series = "CTIT Technical Report Series",
publisher = "Centre for Telematics and Information Technology (CTIT)",
number = "LNCS4549/TR-CTIT-07-72",
address = "Netherlands",

}

Poppe, RW 2007, Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets. CTIT Technical Report Series, no. LNCS4549/TR-CTIT-07-72, Centre for Telematics and Information Technology (CTIT), Enschede.

Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets. / Poppe, Ronald Walter.

Enschede : Centre for Telematics and Information Technology (CTIT), 2007. 9 p. (CTIT Technical Report Series; No. LNCS4549/TR-CTIT-07-72).

Research output: Book/Report › Report › Professional

TY - BOOK

T1 - Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets

AU - Poppe, Ronald Walter

N1 - This text has been accepted for the Computer Vision and Pattern Recognition (CVPR’07) workshop on Evaluation of Articulated Human Motion and Pose Estimation EHuM2). This version of the document is for documentation purposes.

PY - 2007/10

Y1 - 2007/10

N2 - We present an example-based approach to pose recovery, using histograms of oriented gradients as image descriptors. Tests on the HumanEva-I and HumanEva-II data sets provide us insight into the strengths and limitations of an example-based approach. We report mean relative 3D errors of approximately 65 mm per joint on HumanEva-I, and 175 mm on HumanEva-II. We discuss our results using single and multiple views. Also, we perform experiments to assess the algorithm’s generalization to unseen subjects, actions and viewpoints. We plan to incorporate the temporal aspect of human motion analysis to reduce orientation ambiguities, and increase the pose recovery accuracy.

AB - We present an example-based approach to pose recovery, using histograms of oriented gradients as image descriptors. Tests on the HumanEva-I and HumanEva-II data sets provide us insight into the strengths and limitations of an example-based approach. We report mean relative 3D errors of approximately 65 mm per joint on HumanEva-I, and 175 mm on HumanEva-II. We discuss our results using single and multiple views. Also, we perform experiments to assess the algorithm’s generalization to unseen subjects, actions and viewpoints. We plan to incorporate the temporal aspect of human motion analysis to reduce orientation ambiguities, and increase the pose recovery accuracy.

KW - METIS-242003

KW - EC Grant Agreement nr.: FP6/033812

KW - IR-64422

KW - EWI-11258

KW - HMI-CI: Computational Intelligence

KW - HMI-HF: Human Factors

KW - HMI-VRG: Virtual Reality and Graphics

M3 - Report

T3 - CTIT Technical Report Series

BT - Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets

PB - Centre for Telematics and Information Technology (CTIT)

CY - Enschede

ER -

Poppe RW. Evaluating Example-based Pose Estimation: Experiments on the HumanEva Sets. Enschede: Centre for Telematics and Information Technology (CTIT), 2007. 9 p. (CTIT Technical Report Series; LNCS4549/TR-CTIT-07-72).

Access to Document