Fans of Stanley Kubrick‘s 1968 epic 2001: A Space Odyssey will recognize the scenes and motifs in the video below, but get ready to have your mind bent when you learn how they were created. Artist Bhautik Joshi‘s short animation looks like just another artist recreation of a classic, but the reality is so much more woven into the very ideas Kubrick presented in the ’60s than you might think. Watch it now.
2001: A Picasso Odyssey from Bhautik Joshi on Vimeo.
Joshi created the video that transforms Kubrick’s scenes into vivid cubist images using deep neural network based style transfer. The technique requires the application of deep learning algorithms that examine visual data (aka a picture) and morph the individual elements of that picture from one style of pixel to another, effectively taking an image that is styled in one way and recreating it anew. Joshi took film, applied the algorithm, and rendered all the visual data to look like Picasso had painted the scenes from start to finish.
On the one hand, this is unbelievably cool, and just learning about the kind of algorithms that can do this makes my brain stretch and grow. On the other hand, we are teaching software to learn to do things we can do with our own two hands, and any fan of 2001 knows that bad things happen when we give the machines brains!
Deep learning algorithms are not necessarily the first steps towards HAL levels of artificial intelligence, but the perfection of this application of technological advancement is too good to ignore. According to The Epoch Times, the particular type of algorithm that can paint images in different styles has been in existence for less than a year, and it’s already being used by companies like Google in their image recognition practices. Joshi is using them to transform classics of science fiction, and if you take a moment to enjoy his Van Gogh treatment of Blade Runner, you will find the effect absolutely stunning.
What’s the next film you’d like to see him take a crack at with these software paint brushes? Tell us in the comments below!
Images & Video via Bhautik JoshiÂ
