Blood vessel detection from Retinal fundas images using GIFKCN classifier.

In this paper, an automated method for blood vessel detection from retinal fundus images is proposed. Initially, the method extracts the green layer of the retinal image as it contains the requisite information. The noise is removed using a noise removal filter this step will lead to the reduction of artifacts in the final results. The image features are highlighted through morphological operations i.e., top hat and bottom hat transformation on the preprocessed image. The resultant image is then clustered into two clusters representing vessel and non-vessel using generalized improved Fuzzy Kohonen Clustering Network (GIFKCN). The experiments are performed on the Digital Retinal Images for Vessel Extraction (DRIVE) database. The effectiveness of the proposed technique is analyzed using performance metrics like accuracy and sensitivity. A comparative study of the proposed technique with some of the well-known existing methods is done. The comparison shows that the proposed method has improved significantly. [ABSTRACT FROM AUTHOR]