Segmentation of retinal vasculature using phase congruency and hierarchical clustering

Extraction of blood vessels in the fundus images having bright lesions is difficult. A novel approach for the automated segmentation of the vasculature in fundus images with bright lesions is presented in this paper. In this approach, the intensity information from red and green channels of the same fundus image is used to modify the non-uniform illumination in fundus images. Phase Congruency is employed to improve the contrast of vessel segments against the retinal background. The contrast enhanced vessels are then segmented by using hierarchical clustering based histogram thresholding. The experimental study has proved that the proposed algorithm demonstrates better performance than the other vessel detection algorithms reported in the recent literature.