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

Neural network-based image reconstruction for positron emission tomography.

Partha Pratim Mondal1, Kanhirodan Rajan

  • 1Department of Physics, Indian Institute of Science, Bangalore 560012, India. partha@physics.iisc.ernet.in

Applied Optics
|October 29, 2005
PubMed
Summary

A novel Hebbian learning scheme enhances Positron Emission Tomography (PET) image reconstruction by adaptively adjusting pixel interactions. This method produces artifact-free, edge-preserving images superior to traditional Maximum Likelihood and Maximum A Posteriori algorithms.

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

Robust Mixed-Ligand Ni(II)-framework With Pore-Engineered Lewis Acid-Base Sites for Size-selective and One-Pot Mild Synthesis of Bioactive 4H-Pyrans.

Chemistry, an Asian journal·2026
Same author

Single molecule super resolution microscopy reveals formation of Dengue NS2B3 cluster on mitochondrial network and its effect on fragmentation.

Npj viruses·2026
Same author

Unlocking Synergistic Ligand-Metal Interplay in Dual Redox-Active Metal-Organic Framework for High-Efficiency and Durable Overall Water Splitting.

ACS applied materials & interfaces·2026
Same author

Planar lightsheet optical tweezer pLOT for 2D trapping and imaging of freely-moving live cells.

Communications biology·2025
Same author

Role of the Chaperone Protein 14-3-3η in Regulation of the Infection Dynamics of the Influenza A (H1N1) Virus.

Viruses·2025
Same author

Gallium and germanium leaching from jarosite cake by organic acid: a combined experimental and DFT approach.

Journal of environmental management·2025

Area of Science:

  • Medical imaging
  • Computational neuroscience
  • Image processing

Background:

  • Positron Emission Tomography (PET) is a crucial molecular imaging technique.
  • Commonly used reconstruction algorithms like Maximum Likelihood (ML) and Maximum A Posteriori (MAP) have limitations.
  • ML algorithms introduce noise, while MAP algorithms can oversmooth images, especially across edges, and fail to adapt to pixel density variations.

Purpose of the Study:

  • To develop an artifact-free and edge-preserving image reconstruction method for PET.
  • To address the limitations of conventional ML and MAP algorithms in handling interpixel interactions and density variations.
  • To improve the quality of reconstructed PET images by modeling local correlations effectively.

Main Methods:

  • A Hebbian neural learning scheme was proposed to model interpixel interactions.

Related Experiment Videos

  • The approach modifies interaction weights adaptively based on local pixel correlations.
  • This method aims to avoid the oversmoothing issue often seen in MAP algorithms.

    • Main Results:

    • The Hebbian learning-based approach successfully reconstructs PET images with preserved edges and reduced artifacts.
    • Quantitative analysis demonstrated superior performance compared to conventional ML and MAP algorithms.
    • The adaptive weight adjustment effectively models the strength of interpixel interactions.

    Conclusions:

    • The proposed Hebbian learning scheme offers a significant improvement for PET image reconstruction.
    • This method provides a robust way to generate high-quality, artifact-free images by intelligently handling pixel correlations.
    • It represents a promising advancement in molecular imaging reconstruction techniques.