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

Real-time tracking using trust-region methods.

Tyng-Luh Liu1, Hwann-Tzong Chen

  • 1Institute of Information Science, Academia Sinica, Taipei, Taiwan. liutyng@iis.sinica.edu.tw

IEEE Transactions on Pattern Analysis and Machine Intelligence
|September 21, 2004
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

Automatic segmentation of white matter lesions on multi-parametric MRI: convolutional neural network versus vision transformer.

BMC neurology·2025
Same author

Artificial intelligence interpretation of touch print smear cytology of testicular specimen from patients with azoospermia.

Journal of assisted reproduction and genetics·2024
Same author

Early prediction of mortality at sepsis diagnosis time in critically ill patients by using interpretable machine learning.

Journal of clinical monitoring and computing·2023
Same author

ABC-Norm Regularization for Fine-Grained and Long-Tailed Image Classification.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2023
Same author

Toward automatic beam angle selection for pencil-beam scanning proton liver treatments: A deep learning-based approach.

Medical physics·2022
Same author

Determining Carina and Clavicular Distance-Dependent Positioning of Endotracheal Tube in Critically Ill Patients: An Artificial Intelligence-Based Approach.

Biology·2022
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 introduces a novel trust-region framework for real-time object tracking, outperforming traditional line-search methods. The approach effectively integrates color and edge information for robust tracking under rotation and scaling.

Area of Science:

  • Computer Vision
  • Optimization Methods
  • Real-time Systems

Background:

  • Iterative optimization methods for tracking are typically divided into line-search and trust-region techniques.
  • Line-search methods are prevalent in vision applications, while trust-region methods receive less attention.
  • Line-search methods can be viewed as specific instances of trust-region methods.

Purpose of the Study:

  • To establish a flexible and effective trust-region framework for real-time object tracking.
  • To improve tracking performance compared to existing iterative optimization-based trackers.
  • To develop a representation model that handles object rotation and scaling robustly.

Main Methods:

  • Formulation of a trust-region framework for real-time tracking.

Related Experiment Videos

  • Development of a representation model using coupled weighting schemes based on covariance ellipses.
  • Integration of object's color probability distribution and edge density information.
  • Adaptation of various distance functions within the framework.

    • Main Results:

    • The proposed trust-region tracking system demonstrates superior performance compared to line-search-based trackers like mean-shift.
    • The representation model effectively addresses object rotation and nonuniform scaling in a continuous space.
    • Experimental results validate the efficiency and robustness of the proposed tracking framework.

    Conclusions:

    • The trust-region framework offers a more effective approach to real-time object tracking.
    • The integrated color and edge information model enhances tracking accuracy under geometric transformations.
    • The framework's flexibility and demonstrated performance make it a valuable contribution to computer vision.