Artifical intelligence meets art

Abstract

This final year project explores the intersection between Artificial Intelligence (AI), specifically neural network, with Art. This project focuses on neural style transfer, by using the watercolour art style of my original artwork and applying it to a real-life image, producing a result that has a watercolour effect to it while retaining its original details. This is done by using an open-source pre-trained Visual Geometry Group-19 (VGG-19) model and the entire process will be implemented using PyTorch. By combining the content and style representations, this project highlights how AI can be used to produce art. This report will detail the workings of the code used to create the art, its limitation and how it can be resolved. Case studies on artists will be done as well to better understand how AI can be used in art and to influence the art scene. As this has become a bigger topic in recent years, there has been a rise in the concerns regarding the moral use of AI in art which will be explored in this report as well.