Robust Detection and Modeling of the Major Temporal Arcade in Retinal Fundus Images

The Major Temporal Arcade (MTA) is a critical component of the retinal structure that facilitates clinical diagnosis and monitoring various ocular pathologies. Although recent works have addressed quantitative analysis MTA through parametric modeling, their efforts are strongly based on an assumption symmetry in shape. This work presents robust method for detection piecewise modeling fundus images. model consists curve with ability to consider both symmetric asymmetric scenarios. In initial stage, multiple models built from random blood vessel points taken blood-vessel segmented image, following weighted-RANSAC strategy. To choose final model, algorithm extracts width grayscale-intensity features merges them obtain coarse probability function, which used weight percentage inlier each model. procedure promotes selecting high probability. Experimental results public benchmark dataset Digital Retinal Images Vessel Extraction (DRIVE), manual delineations been prepared, indicate proposed outperforms existing approaches balanced Accuracy 0.7067, Mean Distance Closest Point 7.40 pixels, Hausdorff 27.96 while demonstrating competitive terms execution time (9.93 s per image).