Detection and localization of main structures in the retinal images, like blood vessels and Optic Disk (OD), play important role in diagnosis of some diseases like Diabetic Retinopathy (DR). This paper presents a new procedure for automatic extraction of the blood vessels and of the OD in Fundus Fluorescein Angiogram (FFA). The algorithm of vessel extraction starts with the analysis of directional images resulting from sub-bands of curvelet transform in the similar direction and different scales, in order to extract blood vessel centerlines. The final vessel segmentation is obtained using a simple region growing algorithm iteratively, which uses the images resulting from morphological bit plane slicing as aggregated image and vessel centerlines as seeds. Also, the proposed algorithm for OD boundary extraction contains the following stages: At first stage, the FFA images are processed by a sequence of the dilation and erosion operators using multi-structuring elements, in order to remove blood vessels from the FFA images. At second stage, Fast Discreet Curvelet Transform (FDCT) of the image resulting from the previous step is calculated and curvelet coefficients are modified using an exponential function. Then, the image reconstructed from modified curvelet coefficients are used to extract OD regions candidate by applying canny edge detector and morphological operators. At third stage, the information of the blood vessels surrounding the OD region is used for extracting actual location of OD. For this purpose, the proposed method for detecting vessels is used for getting information of vessels around OD candidate regions. Finally, OD boundary is detected by applying distance regularized level set evolution (DRLSE) The proposed method for vessel extraction is tested on the FFA images from angiography unit of Isfahan Feiz hospital, containing 70 FFA images from different DR stages, and on the color images of DRIVE database. The experimental results show accuracy more than 93% and 94% on FFA and color images, respectively. Also, the proposed method for OD boundary extraction is tested on the FFA images from the angiography unit of Isfahan Feiz hospital. The performance of propsed method shows accuracy more than 90%. Keywords: Fundus Fluorescein Angiography, Curvelet Transform, Distance Regularized Level Set Evolution (DRLSE).
