There's an easy and often-used critique of Jackson Pollock’s iconic drip paintings: anyone—even a kid or a monkey—could splatter color on a canvas and make it look like his work. And it's not entirely off base: sure enough, forgers have taken advantage of the difficulty of determining which canvases are authentic and which aren’t. Just last year, one guy got busted for selling around 60 fake Pollocks to collectors for over $1.9 million. Ouch.
But, according to many connoisseurs, critics and fakers don’t give the painter enough credit. There are indeed complexities to Pollock's drip art that show it to be the genuine article. And now there’s a computer program helping to make a science out of the deciphering.
The software uses “computational methods to characterize the low-level numerical differences between original Pollock drip paintings and drip paintings done by others attempting to mimic this signature style,” says Inderscience Publishers. You give it a scan of the possible Pollock, and the program goes to work extracting 4024 numerical image descriptors that the human eye would have trouble deciphering as accurately.
"The human perception of visual art is a complex cognitive task that involves different processing centers in the brain," explained the program’s creator, Lior Shamir. Yet, a computer can pick up on minute pixel-level features of the art we’re just incapable of seeing. The program isn’t perfect—but, reportedly, it has a pretty solid 93 percent accuracy. And it goes a long way in proving that Jackson Pollock’s creations are unique and not just something any shmoe could throw together.
A paper on the software will appear in the upcoming issue of International Journal of Arts and Technology, and Shamir has made his program available to anyone interested. Perhaps it will help save a few collectors some big money: in 2006, an authentic “No.5 1948” was sold to a private enthusiast for a staggering $140 million.