Landmark Detection in Retinal Fundus Images

Abstract

Fundus imaging is one of the most frequent used techniques for screening, diagnosis of eye diseases and abnormalities related to vessels. Optic disc, blood vessels are the main landmarks of the fundus retinal images. If the disease is detected prior then the visual loss can be prevented. Therefore automatic evaluation of fundus images is very necessary and can be used to support immediate diagnosis in ophthalmology as the shape and size of blood vessel and optic disc is an essential indicator of various eye diseases such as diabetic retinopathy and glaucoma. The basis of many automatic evaluations or diagnosis is the segmentation of retina remarkably the detection of optic disc and segmentation of the retinal vessels. The methodology is applied and analysed on available Database Drive. In this work, a method has been presented to segment the optic disc, blood vessels and exudates. Diagnosis of diabetic retinopathy at early stage can be done through detection of blood vessels of retina. Blood vessel segmentation is a helpful tool in the treatment of diabetic retinopathy. The segmentation of retinal vessels is obtained with the help of morphological bit plane slicing from retinal images. The results are compared with standard database and it showed better results in terms of contrast and optic disc and blood vessels are segmented properly