Dataset of Retinal blood Vessels of Healthy and Diabetic Retinopathy Individuals
This dataset contains 50 retinal blood vessels extracted from retinal images available in Diabetic Retinopathy Image Database (DRiDB) public database. The database contains 50 colour fundus images that were captured with Zeiss VISUCAM 200 fundus camera at 45o field of view. 36 of these images has signs of DR while 14 does not show any signs of DR.
Steps to reproduce
A software was developed by the author to automatically detect and extract blood vessels from the retinal images. The raw retinal images were first preprocessed using median filter, mahalanobis distance and Contrast Limited Adaptive Histogram Equalization (CLAHE). Afterwards, D-S uses Laplacian of Gaussian (LoG) and canny edge detection filters were used in detecting the origin of the blood vessel, its edges and its continuous path.