Three-dimensional segmentation of retinal vessels on changes in brightness in OCT images

The veins, the fovea, and the optical disc are three essential features of the human retina. It is challenging to segment and visualizes blood vessels because of the specific conditions that the retinal images have. Some of these terms are especially applicable to the imaging process and imaging modalities, and others are due to the inherent properties of the retina images. Two of the most critical issues affecting image segmentation are lack of proper retina contrast and inappropriate background brightness. Inadequate background image brightness is related to the imaging process and the presence of various veins in the image. Therefore, in this study, an effective method was proposed for extracting and 3D segmentation of blood vessels from retinal images. A multifaceted process for the 3D segmentation of retinal blood vessels in optical coherence tomography slices with fundus ophthalmic images is presented. The proposed algorithm has two distinct clauses, which include 2D segmentation of the retinal blood vessels and three-dimensional segmentation of these vessels based on the calculation of the 2D features of the blood vessels. The proposed algorithm consists of 112 3D OCT spectral scans of the macular eye taken with the Heidelberg HRA imaging system at Tehran Noor Hospital and, the data are obtained from the University of Isfahan database.