Abstract: The purpose of this study was to evaluate the raster photogrammetry based Corneal Topography System by determining: inter-operator variability, reproducibility of images, effects of defocused and decentered images, and the precision of data at different optical zones. 4 human cadaver eyes were used to study the inter-operator variability. To study the reproducibility of images, 3 human cadaver eyes and a test surface doped with flourescine (provided by PAR Vision Systems Corporation) were focused and their images taken four successive times. Defocused and decentered images were taken of 4 human cadaver eyes and four times of the test surface mentioned above. The precision of defocused/decentered cadaver eyes was evaluated at the following optical zones: 3 mm, 4 mm, 5 mm, and 6 mm. All human cadaver eyes used in the above experiments had their epithelial layer removed before imaging. Average inter-operator variability was 0.06 D. In reproducibility attempts, there was an average deviation of 0.28 D for the human cadaver eyes and 0.04 D for the test surface. The defocused and decentered test surface gave an average deviation of 0.09 D. Defocused and decentered cadaver eyes resulted in an average deviation of 0.52 D, 0.37 D, 0.24 D, and 0.22 D at optical zones of 3 mm, 4 mm, 5 mm, and 6 mm, respectively. The imaging method employed by PAR Vision Systems Corporation virtually eliminates inter-operator variability. The PAR Corneal Topography System's clinical usefulness, however, could be improved by improving the reproducibility of images, decreasing the sensitivity to spatial alignment, and increasing accuracy over smaller optical zones.!9
展开▼
