Computers: they make everything so easy these days. They check our spelling for us and help with maths, and now machines have added painting to the list of things they are better than us at.
In a paper entitled, A Neural Algorithm of Artistic Style, and submitted to the Nature Communications journal, researchers showed that a complicated mathematical algorithm (the exact nature of which has not been disclosed) can easily transform a foundation image into the style of a famous artist.
The algorithm forms a “convolutional neural network” (CNN) which, in lay terms, uses object recognition to recreate the foundation image (which can be anything) in the style of a piece of specific art.
Lead author of the paper, University of Tuebingen PhD student Leon Gatys, said:
“The key finding … is that the representations of content [the foundation image] and style [of specific artworks] in the convolutional neural network are separable.
“That is, we can manipulate both representations independently to produce new, perceptually meaningful images.”
The other authors of the paper were Alexander S. Ecker and Matthias Bethge. The paper can be found here (PDF).
This article was written by Hannah Jane Parkinson, for theguardian.com on Wednesday 2nd September 2015 16.11 Europe/Londonguardian.co.uk © Guardian News and Media Limited 2010