Centroid detection in ophthalmic applications

Monochromatic aberrations of the human eye can be objectively measured with an aberroscope or a Hartmann-Shack sensor, which can calculate the wavefront aberrations using displacements in a grid image. These displacements are determined from the centroid locations of the grid points. Often, in optometry practice, the acquired images are of low resolution and low intensity. Current methods for detecting centroids in such images are essentially based on a Canny-Deriche oriented edge-detecting filter. However, its performance has been found to be insufficient in some clinical applications. We propose an alternative method for detecting centroids in grid images based on watershed transformation. The proposed methodology accurately detects the number of grid points for the subsequent estimation of the centroid locations.