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

Assessing visual field clustering schemes using machine learning classifiers in standard perimetry.

Catherine Boden1, Kwokleung Chan, Pamela A Sample

  • 1Department of Ophthalmology, University of California, San Diego, La Jolla, CA 92093-0946, USA.

Investigative Ophthalmology & Visual Science
|December 7, 2007
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

Microbial characterization of oral microbiome in patients with open-angle glaucoma.

Scientific reports·2026
Same author

Diagnostic Performance of Prespecified OCT Rules for Glaucomatous Optic Neuropathy in Nonpathologic Myopia.

JAMA ophthalmology·2026
Same author

Patient- and Family-Centered Care, Communication, and Relationship Development Along a Technological Continuum in the ICU: A Scoping Review.

Critical care explorations·2026
Same author

Progression in Preperimetric Glaucoma Eyes With and Without Microvasculature Dropout.

American journal of ophthalmology·2026
Same author

Ocular Hypertension Treatment Study: Effect of Time After Incident Glaucoma on OCT Optic Nerve Head and Macula Measurements.

American journal of ophthalmology·2026
Same author

Association of Deep Optic Nerve Head Structural Remodeling with Choroidal Microvasculature Dropout in Glaucoma with and without Myopia.

American journal of ophthalmology·2026
Same journal

New Guidelines for Ophthalmic Genetic Studies.

Investigative ophthalmology & visual science·2026
Same journal

Genomic Copy Number Variants Associated With Strabismus and Amblyopia in the All of Us Research Program.

Investigative ophthalmology & visual science·2026
Same journal

Xanthophyll Carotenoid Intake, Plasma Levels, and Retinal Visualization in Aging and Age-Related Macular Degeneration: ALSTAR2.

Investigative ophthalmology & visual science·2026
Same journal

Enhanced Endocytosis and Mitochondrial Stress Underlie Severe Retinitis Pigmentosa With RHO P347L Mutant.

Investigative ophthalmology & visual science·2026
Same journal

Dual-Hit Myopia Mechanism Unveiled by Multi-Omics: Opn1mw Deficiency Primed the Retina for Exaggerated Response to Environmental Defocus.

Investigative ophthalmology & visual science·2026
Same journal

Psychometric Performance of Children With Amblyopia During a Tablet-Based Adaptive Visual Acuity Assessment.

Investigative ophthalmology & visual science·2026
See all related articles

Machine learning classifiers were trained using clustered and unclustered visual field data to detect glaucomatous optic neuropathy (GON). Clustered data did not significantly optimize performance for detecting GON compared to unclustered data.

Area of Science:

  • Ophthalmology
  • Machine Learning
  • Data Science

Background:

  • Glaucomatous optic neuropathy (GON) detection relies on analyzing visual field data.
  • Machine learning (ML) offers potential for improving diagnostic accuracy.
  • Clustering data may enhance ML model performance by organizing complex datasets.

Purpose of the Study:

  • To evaluate if machine learning classifiers trained on clustered visual field data improve the detection of glaucomatous optic neuropathy (GON) compared to unclustered data.
  • To compare the performance of quadratic discriminant analysis (QDA) and support vector machines with Gaussian kernel (SVMg) classifiers under different clustering schemes.

Main Methods:

  • Two ML classifiers (QDA and SVMg) were trained on standard perimetry data from the Diagnostic Innovations in Glaucoma Study (DIGS).

Related Experiment Videos

  • Data was clustered using three schemes, with control conditions of unclustered and randomly assigned data.
  • Classifiers were tested on an independent dataset of early glaucoma and normal subjects.

    • Main Results:

    • Areas under the ROC curve on the training set ranged from 0.85 to 0.92.
    • Clustered data did not significantly optimize sensitivity compared to unclustered data.
    • QDA generally showed higher sensitivity than SVMg, irrespective of clustering method or specificity cutoff.

    Conclusions:

    • Quadratic discriminant analysis (QDA) outperformed SVMg in detecting early glaucoma using visual field data.
    • Data clustering is not essential for visual field data alone but may be beneficial when integrating high-dimensional data, such as structural imaging.