Tracing of Central Serous Retinopathy from Retinal Fundus Images

Abstract

The Fundus images of retina of human eye can provide valuable information about human health. In this respect, one can systematically assess digital retinal photographs, to predict various diseases. This eliminates the need for manual assessment of ophthalmic images in diagnostic practices. This work studies how the changes in the retina caused by Central Serous Retinopathy (CSR) can be detected from color fundus images using image-processing techniques. Localization of the leakage area is usually accomplished by Fluorescein angiography. The proposed work is motivated by severe discomfort occurring to certain patients by the injection of fluorescein dye and also the increase in occurrence of CSR now days. This paper presents a novel segmentation algorithm for automatic detection of leakage site in color fundus images of retina. Wavelet transform method is used for denoising in the preprocessing stage. Contrast Enhancement method is employed for non-uniform illumination compensation and enhancement. The work determines CSR by the localization of leakage pinpoint in terms of pixel co-ordinates and calculates the area of the leakage site from normal color fundus images itself.