Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

3D periodic human motion reconstruction from 2D motion sequences.

Zonghua Zhang1, Nikolaus F Troje

  • 1z.zhang@hw.ac.uk

Neural Computation
|March 27, 2007
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Leveraging ubiquitous mobile sensors to track infant expressions and movements.

Infant behavior & development·2026
Same author

Mechanism-data fusion for quantitative prediction of transmitted wavefront error in multi-point bonded optics.

Optics express·2026
Same author

Individual Brain Charting: fifth release of high-resolution fMRI data for cognitive mapping.

Scientific data·2026
Same author

Motion parallax allows 7-8-month-old infants to distinguish pictures from their referents.

i-Perception·2025
Same author

Model-based 3D reconstruction method with a concave mirror for deflectometry.

Optics express·2025
Same author

Robotic near optical coaxial phase measuring stitching deflectometry for form measurement of specular surfaces.

Optics express·2025

This study reconstructs 3D human motion from 2D data using Fourier decomposition and principal component analysis. The method effectively recovers missing motion data and unknown viewpoints for 3D motion reconstruction.

Area of Science:

  • Computer Vision
  • Human Motion Analysis
  • Biomechanical Engineering

Background:

  • Reconstructing 3D human motion from 2D sequences is challenging due to lost depth information.
  • Periodic human motions, like walking, possess inherent structure that can be exploited for reconstruction.
  • Existing methods often struggle with incomplete data or unknown camera viewpoints.

Purpose of the Study:

  • To develop and evaluate a method for reconstructing 3D periodic human motion from 2D motion sequences.
  • To demonstrate the ability to recover missing motion data and unknown camera viewpoints.
  • To provide a compact and efficient representation for human motion.

Main Methods:

  • Fourier decomposition to create a compact representation of 3D periodic human motion.

Related Experiment Videos

  • Principal Components Analysis (PCA) to learn a low-dimensional linear motion model from 3D Fourier representations.
  • Projection of 2D motion data onto the learned model using least-squares minimization and Bayesian maximum a posteriori probability estimation.

    • Main Results:

    • Successfully reconstructed 3D periodic human motion from 2D sequences.
    • Demonstrated accurate recovery of missing motion data.
    • Showcased the ability to retrieve unknown horizontal viewpoints from 2D motion data.
    • Validated the effectiveness of both least-squares and Bayesian approaches.

    Conclusions:

    • The proposed linear motion model effectively reconstructs 3D human motion from 2D data.
    • The method is robust to missing data and can infer unknown camera parameters.
    • This approach offers a promising solution for 3D human motion capture and analysis from limited 2D observations.