Denis Shiryaev is an AI wizard who has liberally applied his magic to old film—upscaling, colorizing, and otherwise modernizing scenes from Victorian England, late Tsarist Russia, and Belle Époque Paris. He trained machines to restore the earliest known motion picture, 1888’s Roundhay Garden Scene and one of the most mythologized works of early cinema, the Lumière Brothers 50-second Arrival of a Train at La Ciotat Station.
Shiryaev’s casual distribution of these efforts on YouTube can make us take for granted just how extraordinary they are. Such recreations would have been impossible just a decade or so ago. But we should not see these as historic restorations. The software Shiryaev uses fills in gaps between the frames, allowing him to upscale the frame rate and make more naturistic-looking images. This often comes at a cost. As Ted Mills wrote in an earlier Open Culture post on Shiryaev’s methods, “there are a lot of artifacts, squooshy, morphing moments where the neural network can’t figure things out.”
But it’s an evolving technology. Unlike wizards of old, Shiryaev happily reveals his trade secrets so enterprising coders can give it a try themselves, if they’ve got the budget. In his latest video, above, he plugs the NVIDIA Quadro RTX 6000, a $4,000 graphics card (and does some griping about rights issues), before getting to the fun stuff. Rather than make old film look new, he’s “applied a bunch of different neural networks in an attempt to generate realistic faces of people from famous paintings.”
These are, Shiryaev emphasizes, “estimations,” not historical recreations of the faces behind Leonardo’s Mona Lisa and Lady with an Ermine, Botticelli’s model for The Birth of Venus, Vermeer’s for Girl with a Pearl Earring, or Rembrandt’s The Night Watch. In the case of American Gothic, we have a photo of the model, artist Grant Wood’s sister, to compare to the AI’s version. Frida Kahlo's Self-Portrait with Thorn Necklace and Hummingbird gets the treatment. She left perhaps a few hundred photographs and some films that probably look more like her than the AI version.
The GIF-like “transformations,” as they might be called, may remind us of a less fun use of such technology: AI’s ability to create realistic faces of people who don’t exist for devious purposes and to make “deep fake” videos of those who do. But that needn’t take away from the fact that it’s pretty cool to see Botticelli’s Venus, or a simulation of her anyway, smile and blink at us from a distance of over 500 years.