Right on the Monet: Creating Monet-style Art using CycleGAN

Jason Dolorso
Asian Institute of Management

Executive Summary

Generative Adversarial Networks (GANs) is one of the breakthroughs in the field of Neural Networks since it was introduced in 2014 and Image-to-Image translation is one of it's widely known use where an image is translated into another such as turning a horse into a zebra or summer landscape into a winter setting. CycleGANs, a variant of GANs solves one of the difficulties of Image-to-image translation as it eliminates the need for paired samples. This project shows how CycleGAN works and how we can apply it to turn random landscapes into a Monet-like painting. Although the Monet-fied results are satisfying according to groups of people whom the images are shown to, the validation remains the difficult part in the GAN creation process especially with an art-related application due to its subjectivity.

Keywords

generative Adversarial networks, cyclegan, art generation

---

Full text article and source codes can be provided upon request.