Observation Model for Retinal Image Normalization

Abstract

Retinal images are often acquired to diagnose diseases like diabetes, hypertension, and glaucoma. In acquisition process images are non-uniformly illuminated and exhibit local luminosity and contrast drift. This may affect diagnostic process and results in deriving diagnostic parameters. We proposed method that estimates background of retinal images by sub sampling and interpolation. This method based on estimation of luminosity and contrast variability in the background part of image and the subsequent compensation of this variability in whole image. Experimental results show that this method can effectively achieve non uniform illumination and contrast normalization.