Machine Intelligence for Identification of Endothelial Corneal Layer Diseases with Novel Morphology Algorithm

Abstract

Cornea is the sensitive membrane in the eye. The transparent surface of the cornea allows visual data to process into retina. Visual imperfection is due to the functional disorder of the five layers. The vision imperfections are due to either dystrophy or degenerations. This paper focused on endothelium layer is sensitive and good transparent surface. Cell density will play a vital role for diagnosis and to improve the visibility of the objects. In-order to estimate the cell density the image acquired from the cornea is preprocessed for eliminating the noise, to detect the cells and to estimate the cell density. Coherent microscope is used to harvest the cornea images. The median filter is a adopted to curtail the unwanted high frequency components i.e. noise signal. The resulted image is much smoother for further processing. A novel morphology Method has been proposed for analyzing the filtered image for tracing the abnormality in the cell structure. The estimated results with four datasets with morphology method compared with the manually estimated results. This approach exhibits significant superior results for diagnosing the dystrophies.