MULTI-RESOLUTION SEGMENTATION OF RETINAL VESSELS USING OPTIMAL THRESHOLDING METHOD

This paper presents a method for segmenting the vascular structures of retinal images. Retinal vessel segmentation is an essential step of the diagnoses of various retinopathies through the measure of tortuosity of the retinal vessels. Therefore, it is very important to accurately extract blood vessels. The paper analyzes the characteristic of retinal images, secondly gray mathematical morphology theories are used to smooth and strengthen retinal images in order to remove the background and enhance the brightness of retinal blood vessels. Candidate window is selected by Eigen decomposing, and local intensity threshold is applied to the selected candidate window in order to segment the vessels. The tracking strategy is then used to extract thin vessels without being corrupted by noise. Using this multi-resolution segmentation scheme, vessels with different widths can be segmented at different resolutions, although the window size is fixed. Then the tortuosity of the retinal vessels, thus segmented is determined using the curve fitting and deviation analysis.