Glaucoma is an incurable disease that leads to vision loss. Early diagnosis and detection of this pathology provide an advantage in slowing the progression of glaucoma. Accurate segmentation of the optic disc (OD) and optic cup (OC) is useful for diagnosing glaucoma. In recent years, deep learning has achieved remarkable results in the segmentation of OD and OC. However, OC segmentation is more challenging than OD segmentation due to the high variability in shape and unclear boundaries, which reduces the performance of models for detecting and segmenting the OC. This study examines methods for processing fundus images to identify glaucoma pathology, localize it, and analyze its changes over time, which is crucial for accurate diagnosis. The method of fundus image registration and mathematical models play an important role in describing fundus images. The image space model, in particular, provides powerful analytical tools for studying these images. Further image processing is based on neighboring pixels, allowing precise transformation of images based on brightness values in localized areas.