Unsupervised automated retinal vessel segmentation based on Radon line detector and morphological reconstruction

Retinal blood vessel segmentation and analysis is critical for the computer-aided diagnosis of different diseases such as diabetic retinopathy. This study presents an automated unsupervised method segmenting retinal vasculature based on hybrid methods. The algorithm initially applies a preprocessing step using morphological operators to enhance tree structure against non-uniform image background. main processing Radon transform overlapping windows, followed by validation, refinement reconstruction achieve final segmentation. was tested three publicly available datasets local database comprising total 188 images. Segmentation performance evaluated measures: accuracy, receiver operating characteristic (ROC) analysis, structural similarity index. ROC resulted in area under curve values 97.39%, 97.01%, 97.12%, DRIVE, STARE, CHASE-DB1, respectively. Also, results accuracy were 0.9688, 0.9646, 0.9475 same datasets. Finally, average index computed all four datasets, with 0.9650 (DRIVE), 0.9641 (STARE), 0.9625 (CHASE-DB1). These compare best published date, exceeding their several datasets; similar found accuracy.