Abstract:
Understanding the relationship between vascular changes and many diseases requires complex network topology estimation. Ophthalmologists can diagnose and treat eye disease by automatically classifying retinal vascular trees into arteries and veins.
Their projective ambiguity and subtle changes in appearance, contrast, and geometry during imaging make it difficult. We propose a new method to distinguish arteries from veins (A/V) in retinal color fundus images using vascular network topological properties.
We use dominant set clustering to formalize retinal blood vessel topology estimation and A/V classification as pairwise clustering problems. Image segmentation, skeletonization, and significant nodes create the graph.
Edge weight is the inverse Euclidean distance between its end points in the feature space of intensity, orientation, curvature, diameter, and entropy. Intensity and morphology classify the reconstructed vascular network into arteries and veins.
The proposed method was applied to five public databases—INSPIRE, IOSTAR, VICAVR, DRIVE, and WIDE—and yielded high accuracies of 95.1%, 94.2%, 93.8%, 91.1%, and 91.0%. We also released manual annotations of blood vessel topologies for INSPIRE, IOSTAR, VICAVR, and DRIVE datasets to help researchers.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
