Modeling the Optical Properties of the Aging Human Crystalline Lens from Computer Processed Scheimpflug Images in Relation to the Lens Paradox

Of the many methods that have been developed (e.g., phakometry, NMI, etc.) for non-invasive measurement of the geometry of the anterior segment, at present Scheimpflug photography offers the best resolution and the highest accuracy. The primary obstacle encountered with this or any other image based method has been obtaining quantitative measurements of the position and curvature of lens surfaces and zone boundaries from the images directly. Image enhancement (Sobel gradient scanning), and pattern recognition methods (Hough transformation and recursive least squares algorithms) have been applied successfully to this problem. These techniques have been described previously [1] as well as the algorithms used to correct for nonuniform Scheimpflug camera scaling. Combined, these techniques yield representations of lens geometry having sufficient accuracy that ray tracing can be applied to determine lens optical properties from well poised models with one unknown. The image processing methods will be reviewed as an introduction.