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

Using generalized additive models for construction of nonlinear classifiers in computer-aided diagnosis systems.

María J Lado1, Carmen Cadarso-Suárez, Javier Roca-Pardiñas

  • 1Department of Computer Science, University of Vigo, 32004 Ourense, Spain. mrpepa@uvigo.es

IEEE Transactions on Information Technology in Biomedicine : a Publication of the IEEE Engineering in Medicine and Biology Society
|April 19, 2006
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

Testing Covariates Effects on Bivariate Reference Regions.

Statistics in medicine·2025
Same author

A method for determining groups in cumulative incidence curves in competing risk data.

Biometrical journal. Biometrische Zeitschrift·2024
Same author

Author Correction: Reassessment of fluctuating dental asymmetry in Down syndrome.

Scientific reports·2024
Same author

Quantification of temporomandibular joint space in patients with juvenile idiopathic arthritis assessed by cone beam computerized tomography.

Orthodontics & craniofacial research·2023
Same author

Multivariate reference and tolerance regions based on conditional transformation models: Application to glycemic markers.

Biometrical journal. Biometrische Zeitschrift·2023
Same author

Author Correction: A cross-country study on the impact of governmental responses to the COVID-19 pandemic on perinatal mental health.

Scientific reports·2023

Generalized additive models (GAMs) offer an improved alternative to linear discriminant analysis (LDA) for computer-aided diagnosis (CAD) systems. GAMs enhance diagnostic performance by reducing false positives in microcalcification detection.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence in Medicine
  • Statistical Modeling

Background:

  • Computer-aided diagnosis (CAD) systems assist radiologists by acting as a second reader for image interpretation.
  • Evaluating diagnostic performance, often using receiver operating characteristic (ROC) curves, is crucial for CAD systems.
  • False positive reduction and lesion classification are key steps in CAD, frequently employing statistical methods like linear discriminant analysis (LDA).

Purpose of the Study:

  • To introduce and evaluate a novel approach using generalized additive models (GAMs) as an alternative to LDA in CAD systems.
  • To address the limitations of LDA in handling diverse variable types.
  • To improve the diagnostic performance of CAD systems, specifically in reducing false positives.

Main Methods:

Related Experiment Videos

  • Development of a novel approach based on generalized additive models (GAMs).
  • Application of GAM techniques for reducing false detections in a computerized method for detecting clustered microcalcifications.
  • Comparison of GAM performance against the traditional linear discriminant analysis (LDA) model.

Main Results:

  • Linear discriminant analysis (LDA) achieved a sensitivity of 80.52% with 1.90 false positives per image.
  • Generalized additive models (GAMs) improved sensitivity to 83.12% while reducing the false-positive rate to 1.46 per image.
  • GAMs demonstrated superior performance over LDA in the context of clustered microcalcification detection.

Conclusions:

  • Generalized additive models (GAMs) provide a more effective statistical approach compared to LDA for CAD systems.
  • GAMs offer advantages in handling various data types, leading to enhanced diagnostic accuracy and reduced false positives.
  • The study highlights the potential of GAMs to significantly improve the performance of CAD systems in medical image analysis.