We have proposed a new automatic visualization procedure based the ratio of optical densities (ODs) obtained at two different wavelength for the oxygen saturation imaging in human retinal vessels. This method utilized the morphological processing and the line convergence index filter to estimate the reflection image of outside vessels and extract the vessel structure from retinal image, respectively. In the experimental measurement, clear difference between retinal arteries and veins has been observed. In this study, the data processing technique of the line convergence index filter was applied to a color fundus image to investigate the ability of vessel extraction. In addition, four-wavelength imaging was proposed to evaluate oxygen saturation of the retinal capillary vessels and to decrease the influence of the melanin pigmentation.