Creating Abstract Art with Neural Style Transfer

Deep learning algorithms have been used in numerous image processing areas, mainly supervised and unsupervised classification and segmentation applications in a variety of fields. Latest developments in DL algorithms explore generative styles of implementation, such as style transfer of images, general adversarial networks (GANs), and so on. This poster illustrates an application of artistic neural style transfer where a simple set of doodles fed into a deep learning algorithm can create complex patterns of abstract art. It explores a specific style of abstract art called 'Zentangles' where structured patterns are used to create intricate art.