Basic macrotextures structure detection in corneal images using fuzzy techniques

Endothelium analysis of donated corneas used for transplantation is of fundamental importance to determine if the cornea is able to maintain its transparency at the host bed. In this work, we present a segmentation method for analyzing the patterns of the endothelium cells from donated cornea to supply important information for a report of the cornea for transplant. The system is based on fuzzy techniques, what allows taking into account the problems due to little brightness, non-uniform illumination and poor contrast, which are responsible for the cells to be almost imperceptible in certain areas. Moreover, the system adapts to the inherent feature variability of natural textures.