Fig. 4From: Machine learning helps improve diagnostic ability of subclinical keratoconus using Scheimpflug and OCT imaging modalitiesFisher’s score of each variable of different classifiers to discriminate KC eyes from normal eyes. For KC eyes, using the Pentacam HR system alone and combining it with UHR-OCT, the feature that contributed to discrimination most by ranking was the maximum elevation value in the 5 mm area (a, c). Using UHR-OCT alone, the feature that contributed to discrimination most by ranking was SEI (I/S) (b). KC: keratoconus. UHR-OCT: Ultra-high-resolution optical coherence tomography; SEI (I/S): Localized thinning in the vertical meridian in the stromaBack to article page