Determination of the unit size of the corneal endothelial cell mosaic from Fourier component image analysis

Scanning electron micrographs were taken of the central region of the corneal endothelium of cows (Holstein). The cells were outlined and an image overlay generated of approximately 100 cells. Via a Windows-based scanning system, the overlay was subjected to a two-dimensional Fourier transform on a Unix-based system. A custom algorithm was developed (IRIS) to sequentially analyse the Fourier transform pattern. The transform and resultant harmonics spectrum were compared to those obtained from artificial cell mosaics generated from uniform sized symmetrical hexagons. The position, width, and height of the 1st harmonic component of the frequency distribution appears to be derived from average cell-cell border distances across the image. The radial position of the 1st harmonic is inversely related to the unit cell size in the mosaic, i.e. the dominant cell size. The application of such techniques to cell mosaic analyses is discussed.