A blood vessel in the retinal is one of the important organs especially to diagnose diseases such as diabetic retinopathy and glaucoma. In this study, a new method for automatic segmentation of blood vessels in retinal images was presented. The proposed method is based on a hybrid combination between Gray-Level and Moment Invariant techniques. There are consists four stages of processing, (1) preprocessing, (2) feature extraction, (3) classification, and (4) post-processing. The proposed method was compared to the Vascular Tree and Morphological method. Based on the objective evaluation, the proposed method successfully achieved a sensitivity of 98.589% and specificity of 55.544% compared to the others.
