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

Feature selection with limited datasets.

M A Kupinski1, M L Giger

  • 1Department of Radiology, The University of Chicago, Illinois 60637, USA.

Medical Physics
|October 27, 1999
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

SU-E-E-03: Ethics and Professionalism Education in Medical Physics: A Needs Assessment Study.

Medical physics·2017
Same author

ADAPTIVE SMALL-ANIMAL SPECT/CT.

IEEE Nuclear Science Symposium conference record. Nuclear Science Symposium·2015
Same author

Automated detection of mass lesions in dedicated breast CT: a preliminary study.

Medical physics·2012
Same author

Radiographic texture analysis of densitometer-generated calcaneus images differentiates postmenopausal women with and without fractures.

Osteoporosis international : a journal established as result of cooperation between the European Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA·2006
Same author

Computerized mass detection for digital breast tomosynthesis directly from the projection images.

Medical physics·2006
Same author

Computerized detection of mass lesions in digital breast tomosynthesis images using two- and three dimensional radial gradient index segmentation.

Technology in cancer research & treatment·2004
Same journal

Correction to "On the shape of the radiation survival curve in tumor spheroids: The role of oxygen heterogeneity".

Medical physics·2026
Same journal

Multi-view constrained semi-supervised vertebra detection for 3D ultrasound spine volume.

Medical physics·2026
Same journal

Accuracy of quantitative <sup>177</sup>Lu SPECT/CT imaging: A systematic review.

Medical physics·2026
Same journal

Physics-constrained dual-domain network for CBCT reconstruction from orthogonal X-rays in gynecologic radiotherapy.

Medical physics·2026
Same journal

Decomposition-based harmonization for quantitative PET imaging across scanners and radiotracers.

Medical physics·2026
Same journal

Development and evaluation of an in vivo dose-based monitoring system for electron FLASH radiation therapy.

Medical physics·2026
See all related articles

Feature selection for computer-aided diagnosis can vary due to performance distributions. This study analyzes optimal feature subset selection and highlights bias risks when using limited datasets for classifier parameter determination.

Area of Science:

  • Medical imaging analysis
  • Machine learning in diagnostics
  • Radiology informatics

Background:

  • Computer-aided diagnosis (CAD) enhances radiologist accuracy by analyzing medical images.
  • CAD systems extract numerous features from suspect regions for classification.
  • Feature selection aims to identify optimal feature subsets for improved classifier performance.

Purpose of the Study:

  • To investigate the variability of "optimal" feature subset selection in CAD systems.
  • To analyze the impact of classifier performance distribution on feature selection accuracy.
  • To identify potential biases in classifier parameter determination with limited data.

Main Methods:

  • Analytical computation of the probability of selecting the true optimal feature subset.

Related Experiment Videos

  • Evaluation of feature selection probability based on dataset size, feature count, and subset size.
  • Assessment of bias introduced by using the same data for feature selection and classifier parameter tuning.

    • Main Results:

    • The probability of selecting the true optimal feature subset depends on dataset size, total features, selected features, and true optimal subset performance.
    • Classifier performance distributions introduce variations in the selected "optimal" feature subsets.
    • Bias is introduced in classifier parameters when determined from the same limited data used for feature selection.

    Conclusions:

    • Understanding feature selection variability is crucial for reliable CAD system development.
    • Careful consideration of dataset size and feature pool is necessary for effective feature selection.
    • Separate datasets for feature selection and classifier training are recommended to mitigate bias.