Eye Fundus Image Processing Using Fuzzy Logic

In this work, an improvement is made to eye fundus images to highlight the characteristics associated either with diabetic retinopathy and glaucoma pathologies, considering observations collected from specialized medical personnel, which are intended to facilitate diagnosis. Migration from the RGB (Red-Green-Blue) color space to the CIE L*a*b* space, defined by the Commission International d’Eclairage (CIE), is proposed for image processing. Given this migration, the intensity levels of pixel illumination are analyzed and modified by employing two fuzzy inference systems (FIS) to improve the image’s contrast and highlight the details associated with these pathologies. The results section presents the relevant data obtained in this proposal. With them, it is demonstrated that applying fuzzy logic in processing eye fundus images is a viable option to support medical personnel in elaborating a diagnosis, considering that fuzzy schemes had been scarcely used for this type of image.