Classifying visual field data

Sterling Hilton, Joanne Katz, Scott Zeger

Research output: Contribution to journalArticlepeer-review

16 Scopus citations


We develop a prediction model for classifying an eye according to glaucoma status based on its visual field. We develop measures of both diffuse and localized defects in the visual held as potential predictors of glaucoma. To identify predictors of abnormal fields, we must describe the variability in the fields of normal eyes, hence we first model the mean, variance and correlation structures of normal fields with use of generalized estimating equations. The best measures of diffuse loss include a field's mean level, contrasts of the upper and lower halves, and contrasts of the nasal and temporal halves. Local loss is measured by the depth, area, volume and location of the field's largest defect. We develop logistic regression models to classify eyes as having glaucoma or not. We present ROC curves of the results that are highly competitive with current clinical methods of classification.

Original languageEnglish (US)
Pages (from-to)1349-1364
Number of pages16
JournalStatistics in Medicine
Issue number13
StatePublished - Jul 15 1996

ASJC Scopus subject areas

  • Epidemiology
  • Statistics and Probability


Dive into the research topics of 'Classifying visual field data'. Together they form a unique fingerprint.

Cite this