about
Artist Style Transfer via Quadratic Potential (arxiv.org)
1 point by pplonski86 on Mar 1, 2019 | hide | past | pdf | discuss on HN

In plain words: A network repaints a photo in a chosen artist's style, trained against a checker network using a smoother scoring rule that keeps the training steady. The team shared stylized samples and the code, but reports no numbers comparing it with usual style-transfer setups.

Abstract · Artist Style Transfer Via Quadratic Potential

In this paper we address the problem of artist style transfer where the painting style of a given artist is applied on a real world photograph. We train our neural networks in adversarial setting via recently introduced quadratic potential divergence for stable learning process. To further improve the quality of generated artist stylized images we also integrate some of the recently introduced deep learning techniques in our method. To our best knowledge this is the first attempt towards artist style transfer via quadratic potential divergence. We provide some stylized image samples in the supplementary material. The source code for experimentation was written in PyTorch and is available online in my GitHub repository.

Rahul Bhalley, Jianlin Su
arXiv:1902.11108 · cs.CV, cs.LG, stat.ML · submitted Feb 14, 2019 · updated Mar 5, 2019
abstract · pdf · html · 8 pages, 3 figures, uses nips_2018.sty, renamed the network to CycleGAN-QP for maintaining consistency with work

add comment on HN