Retinal vessel segmentation using Morphological Angular Scale-Space

In this paper, segmentation of retinal vessel is done using an innovative procedure named as Morphological Angular Scale-Space (MASS). A linear structuring element rotated at different angles determines the connected components and ensuring that connectivity is not lost across the vessels. A scale-space is created by varying the length of the linear structuring element. Evolution to higher scales gradually reduces non-vessel like elements from the processing image with information from the lower scales being utilized to build the image of higher scales. The method, attains lowest mean square error value at a certain scale, which is observed to be better than most of the other morphological techniques of vessel segmentation. A publicly available DRIVE database is used for analysis of the MASS technique as well as to compare with other methods.