Vangobot paints with one eye open. Its a simple fact that has driven his human operators a bit mad over the last year or so. What does that mean? Let’s start with a bit of text from David Hockney’s Wikipedia page:
David Hockney worked with photography, or, more precisely, photocollage. Using varying numbers of small Polaroid snaps or photolab-prints of a single subject Hockney arranged a patchwork to make a composite image. Because these photographs are taken from different perspectives and at slightly different times, the result is work that has an affinity with Cubism, which was one of Hockney’s major aims—discussing the way human vision works. [ Vangobot's David Hockney inspired paintings ]
Vangobot actually only has one eye. He sees a lot like a camera does with one big gaping eyeball set upon its subject. His “eye” can be either a physical camera attached to his robotic elements or a virtual eye in software that studies still photography.
So why is that a problem? I suppose it matters only of one’s artistic aims, but Vangobot is decidedly not a copy machine. And he is not merely aping the motions and habits of any certain painter or style. Vangobot has a distinct and unique painting style both physically and in approach. His style is all his own and that’s one of many of his abilities that makes him artistic.
But let’s go back to Hockney for a moment. Artist Thomas Schmall writes about Hockney’s suggestions that mirrors and lenses came about during the 1400′s, enabling artists to paint things like never before. And yet, the introduction of such technology brought with it baggage — one eyed baggage. Ultimately what we’re faced with is a technological solution to seeing that is, well, flawed insofar as the human experience.
Combine Vangobot’s single eye with advanced mathematics and fast computers and we may have a way out of the predicament. Take for example, 3-D models from a single still image, a Cornell project originally created by Ashutosh Saxena at Stanford [ is Ashutosh at Cornell now? ] Vangobot does not use the Markov Random Field model referenced by the above linked research, but rather something that we call image decomposition … and something that all artists know already, building paintings from the ground up.
Vangobot’s progress in this field of study is decidedly non-linear. Fits and spurts of progress are often followed by weeks of lows and tearing down to start all over again — hey that sounds a lot like an artist. However, in the here and now is something of a high water mark for working depth from paint using a one eyed robot.
More on this subject must wait for a later date but for now you can watch a time lapse video of Vangobot painting “Experimental Landscape”. This is a painting of land’s end on a disturbed sea under an angry sky. The painting is 40″ by 26″ and is painted with three brushes and approximately 42 purpose made paint mixes. It was painted from a reference photo roughly 110×73 pixels in size.
Here’s part one of Hockney’s video: