Artificial Intelligence Filters Make This Russian Film Look Like Something You’ve Never Seen

Dmitry Nikiforov and Aleksei Korneev are the creators of a new Russian film Delete My Photos, a movie that follows a man’s dream of developing a new dating app. And the film’s unique and revolutionary filters? They are completely out of this world.

📷 popsci.com

Their released trailer is completely covered in filters; something that the new age of app users are already used to. Just like a Snapchat filter, this film makes scenes look a bit artificial. The most important scenes were run through a special program, as popsci.com reports:

They ran key scenes through mobile photo editor Prisma, which uses convolutional neural networks to combine photos with artistic styles, modeled on, for instance, Van Gogh and Roy Lichtenstein. The trailer outcome is like a fluttering, painterly lucid dream. “The final result was mind-blowing,” director Nikiforov told Popular Science. “We did not know how it will be and what expect till the end, because nobody’d done it before in cinema.”

📷 popsci.com

📷 popsci.com

The film’s creators used Prisma, with “all but the video editing component, to solve their accessibility problem.” The app describes itself as a service that “transforms your photos into artworks using the styles of famous artists: Munk, Picasso as well as world-famous ornaments and patterns. A unique combination of neural networks and artificial intelligence helps you turn memorable moments into timeless art.”

Prisma has marketed itself mainly towards the Russian market so far. The neural networks used in the app are similar to those used in Google’s dream sequence, and completely digitize the subject in question, as shown in the photos below.

The neural networks were explained by Popular Science:

neural networks are stacked layers of artificial neurons (run on computers) used to process Google Images. To understand how computers dream, we first need to understand how they learn. In basic terms, Google’s programmers teach an ANN what a fork is by showing it millions of pictures of forks, and designating that each one is what a fork looks like. Each of network’s 10–30 layers extracts progressively more complex information from the picture, from edges to shapes to finally the idea of a fork. Eventually, the neural network understands a fork has a handle and two to four times, and if there are any errors, the team corrects what the computer is misreading and tries again.

You can view the full trailer of the Russian film here: