Early Diagnosis of Retinopathy of Prematurity using Image Processing

Retinopathy of prematurity may lead to blindness in neonates. Retinopathy of prematurity (ROP) is responsible for over 30000 blind children worldwide. Retinopathy of prematurity (ROP) mainly observes premature infants with low birth weight and a major cause for childhood blindness due to failure in early-stage detection. The proposed study enhances the feasibility of prediction of the Retinopathy of Prematurity from the retinal fundus image. It includes the retinal fundus image enhancement, retinal vessel segmentation and calculation of retinal vessel tortuosity. Finally, we use Support Vector Machine for classification of ROP. Experiment shows promising results, 92.38 % accuracy, a sensitivity of 93.83%, and a specificity of 91.68%.