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

Design of a high-sensitivity classifier based on a genetic algorithm: application to computer-aided diagnosis

B Sahiner1, H P Chan, N Petrick

  • 1Department of Radiology, University of Michigan, Ann Arbor 48109-0904, USA. berki@umich.edu

Physics in Medicine and Biology
|November 14, 1998
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

Standardization in Quantitative Imaging: A Multicenter Comparison of Radiomic Features from Different Software Packages on Digital Reference Objects and Patient Data Sets.

Tomography (Ann Arbor, Mich.)·2020
Same author

Drivers, challenges and opportunities of forage technology adoption by smallholder cattle households in Cambodia.

Tropical animal health and production·2017
Same author

FOXP3-stained image analysis for follicular lymphoma: Optimal adaptive thresholding with maximal nucleus coverage.

Proceedings of SPIE--the International Society for Optical Engineering·2017
Same author

WE-E-217A-02: Methodologies for Evaluation of Standalone CAD System Performance.

Medical physics·2017
Same author

WE-E-217A-03: Methodologies for Evaluation of Effects of CAD on Users.

Medical physics·2017
Same author

It's all in the hands: peripheral stigmata of infective endocarditis.

QJM : monthly journal of the Association of Physicians·2014

A novel genetic algorithm (GA) method enhances breast lesion classification by prioritizing high sensitivity. This approach effectively identifies benign cases without missing malignancies, improving diagnostic accuracy in computer-aided diagnosis.

Area of Science:

  • Biomedical Engineering
  • Medical Imaging
  • Machine Learning

Background:

  • Accurate breast lesion characterization is crucial for early cancer detection and effective treatment.
  • Traditional classifiers may struggle to balance high sensitivity and specificity, particularly in identifying benign lesions while minimizing false negatives.

Purpose of the Study:

  • To develop and evaluate a genetic algorithm (GA) based feature selection method for high-sensitivity classifiers in breast lesion characterization.
  • To design a classifier that maximizes specificity in the high-sensitivity region of the Receiver Operating Characteristic (ROC) curve.

Main Methods:

  • A GA was employed to select feature combinations based on the ROC partial area index, focusing on high specificity at high sensitivity.
  • A high-sensitivity classifier using Fisher's linear discriminant was trained with GA-selected features.

Related Experiment Videos

  • Texture features from transformed mammographic regions of interest (ROIs) were extracted using spatial grey-level dependence and run-length statistics.

    • Main Results:

    • The GA-based classifier achieved a significantly larger ROC partial area above a true-positive fraction of 0.95 compared to linear discriminant analysis with stepwise feature selection (LDAsfs).
    • The high-sensitivity classifier correctly identified 61% of benign masses, while LDAsfs identified 34%, with no missed malignant masses in either case.
    • The GA-selected features improved classifier performance in the critical high-sensitivity region.

    Conclusions:

    • Feature selection techniques significantly impact computer-aided diagnosis performance.
    • Genetic algorithms offer a valuable tool for designing specialized classifiers, particularly for lesion characterization requiring high sensitivity and specificity.