via Timo: OpenCV Face Detection: Visualized (by Adam Harvey)
Like the 2D Boxcar visuals that bring Genetic Algorithms to life, I can watch this for a long time. It’s a glimpse into vision itself, and where we may never be able to see how humans really see, in 2012 we are finally getting a glimpse at how we are seen.
There are more things to this video than I can articulate quickly, but part of the frame is that the image is Lena Soderberg.
Lena (“Lenna”) appeared as a Playboy centerfold in 1973, which surfaced in an attic in the Bronx a few years later. I came across the image right around the same time a bunch of computer scientists did, and so it is that she became an algorithmic Eve. The industry hit puberty before I did, and compiled code recalls that centerfold with a precision that no human would even aspire to.
Like a lot of folks, I’ve often thought about how what we see determines how we see. And so it is, that a 1973 Playboy model was imprinted in the distributed AI of cameras around the world, at the same time it embedded itself into my six-year-old brain. The effects of all this are impossibly subtle, they’re embedded deep in a neurological stack.
Which is why this video that shows Lena as she is seen by computers does what art is supposed to do. It’s a model of the world, and the artifice of the model reveals the funny parts, the parts that are too embedded in the act of seeing to ever be visible on their own.
