CROSSING THE UNCANNY VALLEY
I’ve written before about the “uncanny valley,” the term used in discussion of technological attempts to simulate the human visage. It refers to the phenomenon where things intended to look human suddenly seem unrealistic as they closely approach a realistic representation of the human.
There was talk this month at SIGGRAPH about Emily, a completely animated character that promises, in the words of creator David Barton, “new levels of believability in computer animation.”
From the linked piece in the Daily Mail:
To create the footage the University of Southern California Institute for Creative Technologies made a a computer generated replica of Emily.
The actress sat inside a sphere of LED lights while she was photographed making 35 different poses.
The light patterns allowed the shine of her skin to be captured independently from her main skin tone, with hundreds of measurements taken for each millimetre.
‘The CG models provide unprecedented detail of natural facial expressions – down to skin pores and fine wrinkles,’ a spokesman said.
This technique has managed to avoid ‘uncanny valley’ – in which an animation looks less realistic as it approaches human likeness.