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

Random Finite Set Based Bayesian Filtering with OpenCL in a Heterogeneous Platform.

Biao Hu1, Uzair Sharif2, Rajat Koner3

  • 1Robotics and Embedded Systems, Technische Universität München, 80333 München, Germany. hub@in.tum.de.

Sensors (Basel, Switzerland)
|April 19, 2017
PubMed
Summary

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

Wave dispersion analysis of porous functionally graded piezoelectric sandwich panels on Kerr substrates.

Scientific reports·2026
Same author

Correction: Licochalcone A mitigates cisplatin-induced nephrotoxicity by inhibiting ferroptosis and mitochondrial dysfunction via the Nrf2/GPX4 axis.

Frontiers in pharmacology·2026
Same author

Dual-mechanism anti-CD73 antibodies CR201 and CR202 targeting distinct domains for cancer immunotherapy.

Frontiers in immunology·2026
Same author

A mechanosensitive lipolytic factor in the bone marrow promotes osteogenesis and lymphopoiesis.

Cell metabolism·2026
Same author

Excessive vigorous exercise impairs cognitive function through a muscle-derived mitochondrial pretender.

Cell metabolism·2026
Same author

Reinforcement learning in linear embedding space unlocks generalizable control across soft robot configurations.

Nature communications·2026
This summary is machine-generated.

This study accelerates pedestrian detection using OpenCL for real-time Bayesian filtering. The developed system achieves significant speedups, enabling efficient pedestrian tracking for various applications.

Area of Science:

  • Computer Vision
  • Real-time Systems
  • Parallel Computing

Background:

  • Traditional filtering methods prioritize performance over computation time.
  • Real-time applications like pedestrian detection require efficient algorithms.
  • Random finite set (RFS) based Bayesian filtering is a key technique.

Purpose of the Study:

  • To investigate OpenCL for accelerating RFS-based Bayesian filtering.
  • To develop a real-time pedestrian tracking system with high accuracy.
  • To address the computational demands of RFS filtering in heterogeneous systems.

Main Methods:

  • Utilized OpenCL to parallelize Bayesian filtering computations.
  • Developed an efficient pedestrian tracking system implementation.
Keywords:
OpenCLrandom finite set Bayesian filteringreal-time execution

Related Experiment Videos

  • Conducted extensive evaluation and profiling to identify and resolve performance bottlenecks.

    • Main Results:

    • Achieved significant video-throughput improvements, averaging 6x (15 fps to 100 fps).
    • Demonstrated an 18x advantage in worst-case frame processing (2 fps to 36 fps).
    • Successfully met real-time constraints for pedestrian tracking applications.

    Conclusions:

    • OpenCL acceleration is effective for RFS-based Bayesian filtering in real-time systems.
    • The developed system offers a balance of tracking accuracy and computational efficiency.
    • The open-source implementation facilitates further research and application development.