DEEP NEURAL NETWORKS CLASSIFICATION OF PAINTING STYLES
In this paper method and architecture of a neural network for solving the problem of classifying styles of paintings by their digital images are presented. As a result of a large number of experiments conducted to find optimal neural network parameters, the final architecture of the model was obtained, which showed an accuracy of 51.5% for 5 closely related classes and 91% for 5 visually different styles.
Keywords: artificial neural networks, deep learning, image classification, dataset
