Enhancing and segmenting retinal vessels and measuring tortuosity using PSO optimization

Tortuosity is one of the micro vascular changes that is observed in many retinopathies. The measurement of retinal vessel tortuosity is a reliable method used as an ophthalmological diagnostic tool. For instance, Retinopathy of Prematurity (ROP) is a condition that can be identified by an increase in vascular dilation and tortuosity. While normal vessels exhibit a straight or gently curved appearance, age or the presence of retinal diseases can cause them to dilate and adopt more complex shapes. Its early detection can prevent the progression of various retinopathies to a critical stage at which a person may become blind. Ophthalmologists rely on the qualitative measure of tortuosity to assess vessel curvature, ranging from mild to extreme. However, subjective interpretations by practitioners can lead to inconsistent results. To establish a correlation between vascular pathology and tortuosity, it is essential to obtain quantitative measurements. The proposed method employs particle swarm optimization (PSO), average filtering and morphological analysis for vessel detection. Tortuosity assessment involves determining the vessel's centerline and measuring its length. The evaluation of retinal blood vessel detection achieved an accuracy of 98.84%. The tortuosity calculation utilized the relative length variation method, which exhibited the strongest correlation with grading.