作者: Michael K Smolek , Stephen D Klyce , Jeffery K Hovis
DOI: 10.1016/S0161-6420(01)00888-0
关键词: Artificial intelligence 、 Medicine 、 Scale (ratio) 、 Contrast (vision) 、 Corneal topography 、 Brightness 、 Hue 、 Pattern recognition 、 Luminance 、 Palette (computing) 、 Optics 、 Standard deviation
摘要: Abstract Purpose To evaluate the American National Standards Institute (ANSI) corneal topography scale (ANSI Z-80.23-1999) and to develop a Universal Standard Scale (USS) based on scientific principles clinical usefulness. Design Retrospective data analysis computer modeling of frequency power distributions. Methods Four variations color scales were constructed using ANSI standard description. The used 21 contour intervals with step sizes 0.5, 1.0, or 1.5 diopters (D). In example 1, hue was varied, whereas saturation brightness held constant. 2, varied systematically, luminance compensate for spectral sensitivity. 3, adjusted improve contrast at center scale. 4, palette generated by varying red, green, blue intensity levels. USS, colors defined hue, brightness, saturation, appearance optimized contrast. interval determined from deviation keratometry 27 normal corneas. range powers found in 388 maps 12 different conditions. Results ANSI-derived difficult interpret because poor inability associate contours specific 0.5 D slightly improved some patterns but reduced useful did not match user's expectations; abnormal appeared shades which is associated normality. readings population 1.59 D. Hence, USS set More than 99.9% within 30 67.5 D, chosen as USS. Conclusions series that lead misinterpretations. alternative would provide uniform presentation, unambiguous interpretation, greater ease comparison among produced topographers.