Authors : G.V. Shrichandran, S. Sathiyamoorthy and P.D. Sheba Kezia Malarchelvi
Abstract: Retinal images provide early signs of diabetic retinopathy, glaucoma and hypertension. These signs can be investigated based on micro aneurysms or smaller vessels. These studies require accurate tracing of retinal vessel structure from fundus images in an automated manner. However, the existing threshold based segmentation encounters great difficulties such as the detected edges are consisted of discrete pixels and may be incomplete or discontinuous and computationally expensive. To solve above problem, Hierarchical Cat Swarm behaviour based Optimization scheme (HCSO) with Mean Shift Clustering (MSC) algorithm is proposed in this study. In diagnosis, the vessel angles and lengths are changed particularly in junctions and it’s detected by using vessel segmentation. Also the bifurcations and crossings are disconnected and the vessel paths are interrupted in retinal image. So, the proposed system focused these kinds of junction problems. Initially, the input image is pre-processed using top hat filtering to enhance the accurate vessel extraction. Then, the geometric structure based features are extracted by using morphological scheme. Here, the junction problem is analyzed through a connectivity kernel. The experimental result shows the proposed work has efficient and effective vessel segmentation and can be useful for image-aided diagnosis systems and further applications.
G.V. Shrichandran, S. Sathiyamoorthy and P.D. Sheba Kezia Malarchelvi, 2019. An Efficient Vessel Segmentation Based on Hierarchical Swarm Optimization Scheme and Mean Shift Clustering with Vessel Connectivities for Retinal Images. International Journal of Electrical and Power Engineering, 13: 36-49.