Jove
Visualize
Contact Us

Related Experiment Videos

Perceptually based FROC analysis.

Rachna Arora1, Harold L Kundel, Craig A Beam

  • 1Biostatistics Core, H. Lee Moffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Tampa, Florida 33612, USA. arorar@moffitt.usf.edu

Academic Radiology
|December 3, 2005
PubMed
Summary

Analyzing free-response operating characteristic (FROC) data is challenging. A new method using visual scan paths for image segmentation offers an alternative approach for FROC analysis in mammography reading experiments.

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

Enhancing Postharvest Quality and Antioxidant Properties of Mango Using Cinnamic Acid-Enriched Xanthan Gum Coatings.

Plant foods for human nutrition (Dordrecht, Netherlands)·2025
Same author

SPIE Medical Imaging 50th anniversary: history of the Image Perception, Observer Performance, and Technology Assessment Conference.

Journal of medical imaging (Bellingham, Wash.)·2022
Same author

Peripheral immune circadian variation, synchronisation and possible dysrhythmia in established type 1 diabetes.

Diabetologia·2021
Same author

Efficacy of GAD-alum immunotherapy associated with HLA-DR3-DQ2 in recently diagnosed type 1 diabetes.

Diabetologia·2020
Same author

Costimulation Blockade Disrupts CD4<sup>+</sup> T Cell Memory Pathways and Uncouples Their Link to Decline in β-Cell Function in Type 1 Diabetes.

Journal of immunology (Baltimore, Md. : 1950)·2020
Same author

Synchronization of the Normal Human Peripheral Immune System: A Comprehensive Circadian Systems Immunology Analysis.

Scientific reports·2020
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

Area of Science:

  • Radiology
  • Medical Imaging Analysis
  • Statistical Modeling

Background:

  • Analyzing free-response operating characteristic (FROC) data, especially with multiple targets or localization requirements, presents significant challenges.
  • Conventional methods often involve image segmentation (e.g., quadrants) and segment-level analysis, with cases nested within.
  • An alternative approach is needed to address the complexities of FROC data interpretation.

Purpose of the Study:

  • To introduce and evaluate a novel method for FROC analysis that utilizes the reader's visual scan path for image segmentation.
  • To apply this new method to data from a mammography reading experiment.
  • To explore the potential of scan path information in improving FROC analysis.

Main Methods:

  • Recorded the gaze scan path of an experienced radiologist reading 40 mammograms for masses and microcalcifications.

Related Experiment Videos

  • Established "perceptual regions" using a clustering algorithm on visual fixations.
  • Combined reader ratings with scan path-derived segmentation to classify regions containing or not containing targets.
  • Analyzed the generated data using conventional maximum-likelihood methods based on the binormal model and tested goodness-of-fit.

    • Main Results:

    • The conventional maximum-likelihood estimation with the binormal model did not fit the observed data (P < .001).
    • Analysis suggested that the binormal assumption may not hold for segments containing targets.
    • A bimodal distribution model might be more appropriate for this type of data.

    Conclusions:

    • The proposed method offers a viable alternative for analyzing FROC experiments.
    • Further development is needed, particularly in extending statistical models (e.g., mixture models) for perceptually based FROC data.
    • Recording scan paths may enhance receiver operating characteristic (ROC) curve analysis by potentially deconvolving interpretation correlations, allowing for conventional statistical methods after conditioning on the scan path.