0
$\begingroup$

I've read the paper A Neural Algorithm of Artistic Style by Gatys et. al. and I find the application of neural style transfer very fun.

I also read that Exploring the structure of a real-time, arbitrary neuralartistic stylization network by Ghiasi et. al. is a more modern approach to NST.

My question is whether the above paper by Ghiasi et. al. is still the state-of-the-art method in NST, or maybe new algorithms perform even more efficiently.

I shall precise that my goal is to deploy some NST algorithm on a web page as a fun project to apply some deep learning and learn about backend-frontend interactions.

$\endgroup$
1
  • $\begingroup$ you can find all the state-of-the-art papers related to all field in machine learning via this community: paperswithcode.com $\endgroup$ – CuCaRot Mar 15 at 4:57
0
$\begingroup$

So for neural style transfer, using the particular method described in Gatys paper, nobody has done better than using VGG net. This is seemingly due to VGGs inherent stability and inability to learn non-robust features of images. More on this here: https://reiinakano.com/2019/06/21/robust-neural-style-transfer.html

That being said, GANs have had huge success in the field of style transfer, getting much better results than the neural style transfer method described in the paper you mentioned. Cyclegan is one of the best in this respect: https://machinelearningmastery.com/what-is-cyclegan/

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.