Retinal blood vessel segmentation using morphological structuring element and entropy thresholding

An automated segmentation method for Retinal blood vessel extraction is developed. The method uses Morphological structuring element and basic filters for preprocessing. The preprocessing step enhances and makes the retinal images most suitable for vessel extraction. Thresholding technique is employed to extract the blood vessels from the background image. This method is evaluated on the two publicly available data bases namely STARE and DRIVE database. This method is compared with the matched filter method for segmentation and kirsch template method. Our method gives appreciable blood vessel extraction as compared with other methods.