Blood Vessel Segmentation Using SOSVM for Fundus Images

Author(s): R. Nivedha, S. Gayathri

Abstract: In this work, we present an extensive description and evaluation of our method for blood vessel   segmentation in fundus images based on a region growing method. Age-related macular degeneration, glaucoma and diabetic retinopathy is a chronic eye disease that leads to Vision loss. As it cannot be cured, detecting the disease in time is important. Customary segmentation priors like a Potts model or total variation usually fail once handling skinny and elongated structures. The major problem in detection is that, there does not exist a great difference both in color and intensity, so segmentation and edge detection becomes tougher. We overcome this issue by employing a segmentation is done using based region growing method for edge detection and the Parameters of the strategy square measure learned mechanically using a structured output support vector machine, a supervised technique wide used for structured prediction during a variety of machine learning applications.