Registration and analysis of retinal images for diagnosis and treatment monitoring

The paper presents in its initial part fundamental mathematical tools for analysis of retinal images using methods of digital signal processing and enhancement and methods for their registration to monitor changes during the treatment. Bloodvessel trees are then extracted to analyse their specific structures to find retina disorders. The proposed graphical user interface is then used for analysis and monitoring of the set of 20 patients observed during the treatment using the optical coherence tomography. Features obtained are then analysed using selected statistical methods to compare features obtained and to propose the method for classification of healthy and diseased patients. The accuracy achieved was 79 %.