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

Generalizing Swendsen-Wang to sampling arbitrary posterior probabilities.

Adrian Barbu1, Song-Chun Zhu

  • 1Department of Computer Science, University of California, Los Angeles, 8125 Math Science Bldg., Los Angeles, CA 90095, USA. abarbu@ucla.edu

IEEE Transactions on Pattern Analysis and Machine Intelligence
|August 27, 2005
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

IntentQA: Intent Question Answering in Videos by Cognitive Context Reasoning.

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

Cross-robot behavior adaptation through intention alignment.

Science robotics·2026
Same author

MSLNet and Perceptual Grouping for Guidewire Segmentation and Localization.

Sensors (Basel, Switzerland)·2025
Same author

M<sup>2</sup> Diffuser: Diffusion-based Trajectory Optimization for Mobile Manipulation in 3D Scenes.

IEEE transactions on pattern analysis and machine intelligence·2025
Same author

Human-level few-shot concept induction through minimax entropy learning.

Science advances·2024
Same author

Training a Two-Layer ReLU Network Analytically.

Sensors (Basel, Switzerland)·2023
Same journal

Relation DETR+: Exploring Explicit Position Relation Prior for Dense Prediction.

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

RBF++: Quantifying and Optimizing Reasoning Boundaries across Measurable and Unmeasurable Capabilities for Chain-of-Thought Reasoning.

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

CAFE: Cross-View Adaptive Fusion and Cluster Center Enhancement for Robust Multi-View Clustering.

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

DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Driving.

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

Ethics-Aware Safe Reinforcement Learning for Rare-Event Risk Control in Interactive Urban Driving.

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

Learning Shape Anchors for Holistic Indoor Scene Understanding.

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

A new graph partitioning algorithm significantly speeds up computer vision tasks like image segmentation and stereo vision. This method is 100-400 times faster than the Gibbs sampler, offering a more efficient solution for complex energy functions.

Area of Science:

  • Computer Vision
  • Graph Theory
  • Machine Learning

Background:

  • Many computer vision tasks are modeled as graph partition problems with energy minimization.
  • Existing methods like Gibbs sampler are general but slow, while Ncut and graph cuts are fast but limited to specific energy functions.

Purpose of the Study:

  • To develop a novel, general, and efficient inference algorithm for graph partition problems in computer vision.
  • To generalize the Swendsen-Wang method for arbitrary probability distributions on graph partitions.

Main Methods:

  • The algorithm computes graph edge weights from local image features.
  • It iteratively performs graph clustering (probabilistic edge cutting) and graph relabeling (simultaneous component flipping).
  • The method simulates ergodic and reversible Markov chain jumps in the space of graph partitions.

Related Experiment Videos

Main Results:

  • The algorithm demonstrates significant speed improvements: 100-400x faster than Gibbs sampler and 20-40x faster than DDMCMC for segmentation.
  • For stereo vision, it achieves results comparable or superior to graph cuts and belief propagation.
  • The algorithm can automatically infer generative models.

Conclusions:

  • The proposed algorithm offers a computationally efficient and broadly applicable solution for graph partition problems in computer vision.
  • It outperforms existing methods in speed and, for stereo vision, in result quality.
  • This work advances inference techniques for energy minimization in vision tasks.