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

Multiple collaborative kernel tracking.

Zhimin Fan1, Ming Yang, Ying Wu

  • 1Electrical Engineering and Computer Science Department, Northwestern University, Evanston, IL 60208, USA. zfa825@ece.northwestern.edu

IEEE Transactions on Pattern Analysis and Machine Intelligence
|May 15, 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

Long-term effects of drug-eluting stents versus bare metal stents on patients with acute ST-elevation myocardial infarction undergoing primary percutaneous coronary intervention: outcomes of 3-year clinical follow-up.

Chinese medical journal·2012
Same author

[Effect of safflor yellow B on vascular endothelial cells injury induced by angiotensin-II].

Yao xue xue bao = Acta pharmaceutica Sinica·2012
Same author

Core-shell hybrid liposomal vesicles loaded with panax notoginsenoside: preparation, characterization and protective effects on global cerebral ischemia/reperfusion injury and acute myocardial ischemia in rats.

International journal of nanomedicine·2012
Same author

Conjugation of cyclodextrin with fullerene as a new class of HCV entry inhibitors.

Bioorganic & medicinal chemistry·2012
Same author

Expression differences of circulating microRNAs in metastatic castration resistant prostate cancer and low-risk, localized prostate cancer.

The Prostate·2012
Same author

[Intestinal absorption effect of Angelica dahurica extract on puerarin of puerariae Lobatae Radix].

Zhongguo Zhong yao za zhi = Zhongguo zhongyao zazhi = China journal of Chinese materia medica·2012
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 addresses unobservable motion parameters in visual dynamic systems by introducing a novel kernel-based tracking approach. The method utilizes redundant local motion parameters for fully observable estimators, enhancing practical visual tracking applications.

Area of Science:

  • Computer Vision
  • Robotics
  • Dynamic Systems

Background:

  • Unobservable motion parameters pose a challenge in visual dynamic systems, limiting tracking accuracy.
  • Traditional methods often rely on compact but uncorrelated global motion parameters, leading to observability issues.
  • Singularity is a critical problem in visual dynamic systems, affecting parameter estimation.

Purpose of the Study:

  • To investigate and address the issue of singularity and unobservable motion parameters in kernel-based tracking.
  • To present a novel motion field representation for designing fully observable kernel-based motion estimators.
  • To enable efficient estimation of high-dimensional motion fields in practical visual tracking scenarios.

Main Methods:

  • A novel motion field representation using redundant, sparsely correlated local motion parameters.

Related Experiment Videos

  • Design of fully observable kernel-based motion estimators based on the proposed representation.
  • Efficient estimation of high-dimensional motion fields through collaboration of local estimators.

    • Main Results:

    • The proposed approach facilitates the design of fully observable kernel-based motion estimators.
    • High-dimensional motion fields can be efficiently estimated by collaborating simpler local estimators.
    • The novel representation overcomes limitations of compact, uncorrelated global motion parameters.

    Conclusions:

    • The developed kernel-based tracking approach effectively handles unobservable motion parameters.
    • Redundant local motion parameter representation enhances the observability of visual dynamic systems.
    • The collaborative estimation strategy ensures practical and efficient high-dimensional motion field estimation.