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

Optimization of wavelet decomposition for image compression and feature preservation.

Shih-Chung B Lo1, Huai Li, Matthew T Freedman

  • 1Center of Imaging Science and Information Systems, Radiology Department, Georgetown University Medical Center, 2115 Wisconsin Avenue. N.W., Suite 603, Washington, D.C. 20007, USA. lo@isis.imac.georgetown.edu

IEEE Transactions on Medical Imaging
|September 6, 2003
PubMed
Summary

A neural network framework optimizes wavelet kernels for image processing, enhancing compression and minimizing errors. Daubechies wavelets excel in general compression, while Haar wavelets are best for sharp edges.

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

MRI radiomics and <sup>90</sup>Y PET dosimetry for predicting hepatocellular carcinoma response after radioembolization.

BMC cancer·2026
Same author

Removal of pentavalent vanadium from water by Fe-Ni loaded multi-walled carbon nanotubes.

Journal of environmental management·2026
Same author

Deep Learning Model With Nodule Indexing Tailored to Early-Stage Lung Cancer Detection.

Journal of the American College of Radiology : JACR·2026
Same author

Editorial: Microbial ecological and biogeochemical processes in the soil-vadose zone-groundwater habitats, volume III.

Frontiers in microbiology·2026
Same author

Ultra-high temperature bacterial agents enhance heavy metal passivation and antibiotic degradation in compost.

Frontiers in microbiology·2025
Same author

ANCA-associated vasculitis combined with coexisting Aspergillus fumigatus and Mycobacterium avium complex infections: a case report.

BMC infectious diseases·2025

Area of Science:

  • Medical Imaging
  • Signal Processing
  • Computer Vision

Background:

  • Wavelet transforms are crucial for image compression and feature extraction.
  • Selecting optimal wavelet kernels is challenging for specific image processing tasks.
  • Existing methods lack a systematic approach to optimize wavelet selection for diverse image types.

Purpose of the Study:

  • To develop a neural-network-based framework for optimizing wavelet kernels in image processing.
  • To identify wavelets that minimize errors and maximize compression efficiency for various image patterns.
  • To evaluate wavelet performance on medical images (mammograms, CT, MRI) and standard test images.

Main Methods:

  • Employed a linear convolution neural network to search for optimal wavelet kernels.

Related Experiment Videos

  • Evaluated wavelet performance using metrics like mean-square-error and compression efficiency.
  • Tested tap-4 wavelets on mammograms, CT head images, MRI, and Lena images.
  • Analyzed wavelet filter spectra to correlate characteristics with compression and feature preservation.

    • Main Results:

    • Daubechies wavelets and similar types offer high compression efficiency and low mean-square-error for general textures and mammographic microcalcifications.
    • Haar wavelets perform best on sharp edges and low-noise smooth regions.
    • A specialized wavelet with specific low-pass filter coefficients demonstrated superior preservation of microcalcification features.

    Conclusions:

    • The developed neural network framework effectively optimizes wavelet kernels for image processing tasks.
    • Wavelet selection significantly impacts compression efficiency and feature preservation, with different wavelets suited for different image characteristics.
    • The optimization approach is generalizable to other wavelet-based image analysis applications.