Automated Extraction and Analysis of Retinal Blood Vessels with Multi Scale Matched Filter

In this paper, an automated approach is proposed to extract and investigate the blood vessel from retinal images using the heuristic algorithm based Multi Scale Matched Filter (MSMF). Retinal blood vessel examination is essential to examine a variety of retinal diseases. In the proposed approach, Social Group Optimization (SGO) based MSMF is designed to support efficient mining of the blood vessels from the retinal image dataset. During this study, the well-known retinal datasets, such as DRIVE and STARE are considered. The aim of the proposed approach is to find out best possible filter values for MSMF to attain enhanced precision during vessel extraction. All the experimentation are carried out using Matlab and well known image similarity measures and the statistical parameters are computed. Finally, the effectiveness of the SGO based MSMF is confirmed against the similar techniques existing in the literature. The results confirms that, proposed approach offers better values of similarity index and image statistical measures compared to the procedures existing in the literature.