Image analysis of the corneal epithelium.

The method for quantifying corneal epithelial images has been examined. When pictures were developed, the images were stored in the memory of the image processing system by video camera. The distribution of the density of the images was analyzed. The density of each pixel of an image was assigned one of 256 grades, from completely black to completely bright. The average (AV), the standard deviation (SD), the coefficient of variation (CV), the skewness (SK) and the sharpness (S) of the density were calculated. Normal epithelium could be differentiated from abnormal by subjectively observing the shapes of the cells, the sizes of the cells and the irregularity of the cell pattern. But there were no statistical differences between the two groups with regard to the AV, SD, CV, SK and S of normal (n = 4) and abnormal (n = 8) corneal epithelial images. Simple analysis of density distribution could not be used as a parameter for the evaluation of the cells.