US RESEARCHERS HAVE TAUGHT ARTIFICIAL INTELLIGENCE TO CONVERT PICTURES IN PHOTOS
Researchers at the University of Berkeley are working on a new photo editor that uses artificial intelligence. Specialists trained neural network to convert paintings by Van Gogh, Monet and other famous artists in realistic images – in fact, this technology is the opposite of popular applications Prisma.
The system can also turn apples into oranges and horses – in zebras.
The software is connected to the base image Flickr, where it is self-learning and the results are checked both manually and automatically. Authors of the project sought two-way conversion unchanged. This means that the application has to issue an initial result of the processing of edited photos back.
For developers who want to try the editor open source project on GitHub – ready for service on the basis of this development is absent.