The “Uncanny Valley” derives from the notion that as computer animations of human appearance, motion and behavior become more and more realistic, there comes a point when the animations are close to realistic but not quite realistic and instead appear “uncanny” or “spooky” or “strange” or “spooky” or “weird” or “erie”.

There is no doubt that this valley exists. The question is how real computer animations of humans need to be to not fall into the uncanny valley. Achieving seemingly-perfect realism in appearance, motion and behavior is computationally expensive. The question is what can we get away with? How computationally spartan can we be and still avoid falling into the valley?

In this talk I discuss elements of realism in human animation, consider example animations on both sides of and in the valley, and discuss several experimental studies designed to answer questions such as: does the uncanny valley exist; what causes “uncanniness”; are some aspects of realism more important than others; are we sensitive to subtle realism cues such as smiles and eye blinks; and what happens if some aspects of animation are very real while others are not?

Bio

Jim Foley is an Emeritus Professor in the College of Computing at Georgia Tech. He is a member of the United States National Academy of Engineering, and a Fellow of AAAS, ACM and IEEE. He has received two lifetime achievement awards, one in computer graphics (the Coons award) from ACM/SIGGRAPH and another in human-computer interaction from ACM/SIGCHI.

Foley is co-author of five books: Fundamentals of Interactive Computer Graphics, three editions of Computer Graphics Principles and Practice, and Introduction to Computer Graphics.

He joined Georgia Tech in 1991 as the founding director of the Graphics, Visualization and Usability Center. In 1996, US News and World Report ranked the Center number one for graduate computer science work in graphics and user interaction. From 1996-99, he was first Director of MERL - Mitsubishi Electric Research Laboratory and them CEO and Chairman of Mitsubishi Electric Information Technology Center America, responsible for Mitsubishi's corporate R&D in North America.

After returning to Georgia Tech, he was chairman (2001-2005) of the Computing Research Association - an organization of over 250 computer science and computer engineering departments, professional societies and research labs. From 1975-76 and 2008-11 he served as Vice President of ACM’s Special Interest Group for Graphics (SIGGRAPH).