Neural Networks with Macrolayers for Classification and Prognosis of Pathology of the Retina of the Eye

Neural networks with macrolayers are suggested to provide prognosis and differential diagnosis of diseases of blood vessels of the ocular fundus. The neural networks contain K groups of informative characters and K + 1 macro-layers. The number of neurons in the first layer of each neural network K is determined by the number of symptoms significant for diagnosis of pathologies of the ocular fundus. Control tests demonstrated that the prognosis and differential diagnosis of thrombosis of the central retinal vein and its branches is appropriate at level 0.9.