Image style recognition using convolutional neural networks

Daniel Kvak


Recent technological advances have made it possible to digitize an increasing number of works of art. It is no wonder, then, that the automatic classification, indexing and retrieval of fine art images has received increased attention in recent years. However, compared to the classification of natural objects, the classification of works of art is a much more challenging task. Convolutional neural networks (CNN), which consist of a number of layers of convolutional matrix operations, are an ideal solution for image analysis and object identification. CNN uses a similar concept to the graphics filters used in graphics editors, but in a much more complex and computationally intensive manner.