Analysis of the Effects of Astigmatism and Misalignment on Corneal Surface Reconstruction From Photokeratoscopic Data
Jianyi Wang, David A. Rice, Stephen D. Klyce
Abstract
Jianyi Wang, David A. Rice, Stephen D. Klyce
Abstract
This article presents a general method of error analysis for reconstruction algorithms employed by corneal topography systems and applies this method to assess the accuracy of the Wang reconstruction algorithms. Using mathematical surface reconstructions, the method detects artifacts introduced in topographic analysis by astigmatism and misalignment. For the Wang algorithms, error introduced by astigmatism was less than 1% of the true value for surfaces with more than 10.00 diopters of astigmatism. A 1-mm translational misalignment of the analyzed surface introduced error of no more than 1.3%. Errors caused by rotational misalignment were less than 3% for misalignments of 5 degrees or less. It is concluded from the error analysis method that the Wang algorithms are sufficiently accurate for use in the analysis of corneal topography. Similar analyses need to be accomplished for the validation of commercially available corneal topography systems.
OpenAlex reports 29 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This article presents a general method of error analysis for reconstruction algorithms employed by corneal topography systems and applies this method to assess the accuracy of the Wang reconstruction algorithms. Using mathematical surface reconstructions, the method detects artifacts introduced in topographic analysis by astigmatism and misalignment. For the Wang algorithms, error introduced by astigmatism was less than 1% of the true value for surfaces with more than 10.00 diopters of astigmatism. A 1-mm translational misalignment of the analyzed surface introduced error of no more than 1.3%. Errors caused by rotational misalignment were less than 3% for misalignments of 5 degrees or less. It is concluded from the error analysis method that the Wang algorithms are sufficiently accurate for use in the analysis of corneal topography. Similar analyses need to be accomplished for the validation of commercially available corneal topography systems.
Key concepts: Astigmatism, Corneal topography, Dioptre, Error analysis, Mean squared prediction error, Optics, Algorithm, Computer science