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

Recovering 3D human body configurations using shape contexts.

Greg Mori1, Jitendra Malik

  • 1School of Computing Science, Simon Fraser University, Burnaby, BC V5A 1S6, Canada. mori@cs.sfu.ca

IEEE Transactions on Pattern Analysis and Machine Intelligence
|June 24, 2006
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

NeuralFeels with neural fields: Visuotactile perception for in-hand manipulation.

Science robotics·2024
Same author

Ego4D: Around the World in 3,600 Hours of Egocentric Video.

IEEE transactions on pattern analysis and machine intelligence·2024
Same author

Real-world humanoid locomotion with reinforcement learning.

Science robotics·2024
Same author

Navigating to objects in the real world.

Science robotics·2023
Same author

Expert-level detection of acute intracranial hemorrhage on head computed tomography using deep learning.

Proceedings of the National Academy of Sciences of the United States of America·2019
Same author

Multi-View Supervision for Single-View Reconstruction via Differentiable Ray Consistency.

IEEE transactions on pattern analysis and machine intelligence·2019
Same journal

HardFlow: Hard-Constrained Sampling for Flow-Matching Models Via Trajectory Optimization.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Industrial Brain: Self-Evolving Neuro-Symbolic Autonomy with Causal Resilience for Cyber-Physical Systems.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Adaptive Hardness-Driven Dictionary Distillation for Incomplete Streaming View Clustering.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Task-KV: Task-aware KV Cache Optimization via Semantic Differentiation of Attention Heads.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Achieving Text-based Person Retrieval with Any Granularity.

IEEE transactions on pattern analysis and machine intelligence·2026
See all related articles

This study estimates 3D human pose from single 2D images by matching input images to a library of exemplar views. The method accurately transfers joint locations for 3D pose estimation, applicable to video sequences.

Area of Science:

  • Computer Vision
  • Human Pose Estimation
  • 3D Reconstruction

Background:

  • Estimating 3D human pose from 2D images is a challenging problem in computer vision.
  • Accurate human pose estimation is crucial for applications in robotics, animation, and surveillance.

Purpose of the Study:

  • To develop a method for estimating 3D human body configuration and pose from a single 2D image.
  • To enable robust human pose estimation by leveraging a library of pre-annotated 2D human body exemplars.

Main Methods:

  • A novel approach using shape context matching and a kinematic chain-based deformation model to match input 2D images to stored 2D exemplars.
  • Transferring manually labeled 2D joint locations from the closest exemplar to the input image.
  • Utilizing an existing algorithm to estimate the 3D body configuration and pose from the identified 2D joint locations.

Related Experiment Videos

Main Results:

  • Successful estimation of 3D human pose from single 2D images across various datasets.
  • Demonstrated the applicability of the technique to video sequences by treating each frame as an independent recognition task.
  • The correspondence process relies on the availability of a sufficiently similar exemplar view in the stored library.

Conclusions:

  • The proposed method provides a viable approach for single-image 3D human pose estimation.
  • The technique shows promise for real-time applications like video analysis and human-computer interaction.
  • Future work could explore reducing the reliance on exact exemplar matches and improving robustness to occlusions.